Power distribution network fault positioning system and method under interconnection perception
By constructing a distribution network fault location system under interconnected sensing, and utilizing deep learning models and explicit data extraction technology, the problems of large data volume, feature redundancy, and noise interference in distribution network fault location are solved, achieving efficient and accurate fault identification and location.
Patent Information
- Application Number
- CN202511476919.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-10-16
- Publication Date
- 2025-12-05
AI Technical Summary
In existing technologies, fault location in distribution networks relies on data collection from a limited number of monitoring points, resulting in incomplete information perception, low efficiency in fault feature extraction, and poor location accuracy. Especially in the case of high-penetration distributed energy access, the fault waveform characteristics exhibit intermittency and nonlinearity, making it difficult for existing methods to achieve accurate and rapid fault identification and location.
By constructing a distribution network fault location system under interconnected sensing, including a fault type identification module, a mapping library construction module, and an explicit data extraction module, a deep learning model is used to identify predictable fault types, a fault type-explicit window mapping library is constructed, explicit operational sensing waveform datasets are extracted, and interconnected sensing amplification processing is performed before finally inputting them into the next-level fault location unit for analysis.
It reduces the amount of data, minimizes feature redundancy and noise interference, improves the accuracy and reliability of fault identification and location, and enables fast and accurate fault location.
Smart Images
Figure CN121069103A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of power distribution network, and particularly relates to a power distribution network fault positioning system and method under interconnected sensing. BACKGROUND
[0002] With the increasing scale of power distribution network, the complexity of its operation state and the fault risk are also improved. Traditional power distribution network fault positioning mainly relies on data collection of limited monitoring points and local feature analysis, and has problems such as incomplete information sensing, low fault feature extraction efficiency, and lag response to early and complex faults. With the access of high penetration rate distributed energy, the fault waveform characteristics show stronger intermittency, uncertainty and nonlinearity, and the existing fault positioning method is difficult to realize accurate and rapid fault identification and positioning. Internet of Things and interconnected sensing provide a new solution for power distribution network state monitoring, but how to effectively extract key fault features from massive initial operation sensing waveform data and identify early fault types becomes a difficult problem to be solved, which further affects the accuracy and reliability of power distribution network fault identification and positioning processing.
[0003] Therefore, in the related art, there are problems of large initial data volume, feature redundancy and noise interference, which result in low fault feature extraction efficiency and poor positioning accuracy. SUMMARY
[0004] The present application provides a power distribution network fault positioning system and method under interconnected sensing, which solves the technical problems of large initial data volume, feature redundancy and noise interference in the prior art, which result in low fault feature extraction efficiency and poor positioning accuracy, and achieves the technical effect of reducing data volume and improving the accuracy of power distribution network fault identification and positioning.
[0005] The application provides a power distribution network fault positioning system under interconnected perception, which comprises a fault type identification module, a mapping library construction module, an explicit data extraction module and a fault positioning result acquisition module.
[0006] In possible implementation manners, the power distribution network fault positioning system under interconnected perception further performs the following processing: collecting a running perception waveform sample set of the target power distribution network under a known fault type label sample; extracting a presequence perception waveform sample set from the running perception waveform sample set, obtaining a presequence perception waveform sample set, and identifying a predictive waveform feature vector sample of the presequence perception waveform sample set; training a predictive fault type identifier according to the predictive waveform feature vector sample and the known fault type label sample, identifying a predictive fault type of the initial running perception waveform data set by using the predictive fault type identifier, and determining a predictive fault type label.
[0007] In possible implementation manners, the power distribution network fault positioning system under interconnected perception further performs the following processing: the known fault type label sample at least comprises a single-phase grounding fault type, an inter-phase short-circuit fault type, a two-phase grounding short-circuit fault type, a three-phase short-circuit fault and a broken line fault; and the running perception waveform sample set comprises running perception waveform samples under at least three different fault levels.
[0008] In a possible implementation, the power distribution network fault location system under interconnected perception further performs the following processing: center perception waveform sample set extraction is performed on the running perception waveform sample set, a center perception waveform sample set is obtained, and a center waveform feature vector sample of the center perception waveform sample set is identified; the center waveform feature vector sample is analyzed, and a key waveform feature corresponding to each known fault type label is determined; a dominant decision rule of the key waveform feature is defined, the center perception waveform sample set is analyzed according to the dominant decision rule, and a dominant time window sample of the key waveform feature is obtained; and a mapping relationship between each known fault type label and the dominant time window sample is fitted, to obtain a fault type-dominant window mapping library.
[0009] In a possible implementation, the power distribution network fault location system under interconnected perception further performs the following processing: a first key waveform feature and a second key waveform feature of a current known fault type label are obtained; a first dominant time window and a second dominant time window are obtained according to the first key waveform feature and the second key waveform feature; and a time window intersection of the first dominant time window and the second dominant time window is taken as a label dominant time window of a current known fault type mapping.
[0010] In a possible implementation, the power distribution network fault location system under interconnected perception further performs the following processing: the dominant decision rule includes a dominant decision triple, the dominant decision triple includes a key waveform feature index, a feature dominant index, and a dominant decision threshold; a feature dominant index of a current key waveform feature is identified according to the dominant decision rule, and a corresponding dominant decision threshold is extracted according to the feature dominant index; the center perception waveform sample set is analyzed based on the feature dominant index, and a dominant perception waveform sample that meets the dominant decision threshold is obtained through binary decision, and a time window corresponding to the dominant perception waveform sample is marked as a dominant time window sample of the current key waveform feature.
[0011] In a possible implementation, the power distribution network fault location system under interconnected perception further performs the following processing: a fault location algorithm type of the next-level fault location unit is read, a waveform input template used for fault location is determined according to the fault location algorithm type; and interconnected perception amplification processing is performed on the dominant running perception waveform data set according to the waveform input template, to obtain an amplified dominant running perception waveform data set.
[0012] The application also provides a power distribution network fault positioning method under interconnected perception, comprising the following steps: extracting an initial operation perception waveform data set of a target power distribution network according to an interconnected perception network, performing predictive fault type identification on the initial operation perception waveform data set, and determining a predictive fault type label; constructing a fault type-explicit window mapping library, including a mapping relationship between a fault type label and an explicit time window, the explicit time window being a time window in which a feature explicitness index of interconnected perception data of the interconnected perception network is greater than a preset feature explicitness index threshold; mapping the predictive fault type label through the fault type-explicit window mapping library, determining a mapped explicit time window for the interconnected perception network, and extracting an explicit operation perception waveform data set under the mapped explicit time window; performing interconnected perception amplification processing on the explicit operation perception waveform data set, obtaining an amplified explicit operation perception waveform data set, inputting the amplified explicit operation perception waveform data set into a next-level fault positioning unit for analysis, and obtaining a fault positioning result.
[0013] The interconnected perception-based power distribution network fault positioning system and method, the fault type identification module, the mapping library construction module, the explicit data extraction module, and the fault positioning result acquisition module are used to extract an initial operation perception waveform data set and perform predictive fault type identification according to an interconnected perception network, construct a fault type-explicit window mapping library, map a predictive fault type label and extract an explicit operation perception waveform data set, perform interconnected perception amplification processing, input an amplified explicit operation perception waveform data set into a next-level fault positioning unit for analysis, and obtain a fault positioning result. The technical problems of large initial data volume, feature redundancy, and noise interference, which result in low fault feature extraction efficiency and poor positioning accuracy, are solved, and the technical effect of reducing data volume and improving power distribution network fault identification and positioning accuracy is achieved. BRIEF DESCRIPTION OF DRAWINGS
[0014] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings of the embodiments of the present disclosure will be briefly introduced below. In the present application, a flowchart is used to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the foregoing or the following operations are not necessarily performed in sequence. On the contrary, various steps can be processed in reverse order or simultaneously as needed. Meanwhile, other operations can be added to these processes, or one or more steps can be removed from these processes.
[0015] Figure 1 The power distribution network fault positioning system under interconnected perception provided by the embodiments of the present application is shown in the structural schematic diagram.
[0016] Figure 2 The power distribution network fault positioning method under interconnected perception provided by the embodiments of the present application is shown in the flowchart.
[0017] Reference signs: fault type identification module 10, mapping library construction module 20, explicit data extraction module 30, fault location result acquisition module 40. DETAILED DESCRIPTION
[0018] The above description is only a summary of the technical scheme of the present application. In order to enable the technical means of the present application to be more clearly understood, and to be implemented according to the content of the description, and in order to enable the above and other purposes, characteristics and advantages of the present application to be more apparent and easy to understand, the following specific embodiments of the present application are described.
[0019] In order to make the purposes, technical schemes and advantages of the present application more clear, the following will further describe the present application in combination with the drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by those skilled in the art without making creative labor are within the scope of protection of the present application.
[0020] In the following description, "some embodiments" are related to a subset of all possible embodiments, but it can be understood that "some embodiments" can be the same subset or different subset of all possible embodiments, and can be combined with each other without conflict. The term "first\second" is only to distinguish similar objects, and does not represent the specific order of the objects. The terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, system, product or server including a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or modules not clearly listed or inherent to these processes, products or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as understood by those skilled in the art in the technical field to which the present application belongs. The terms used herein are only for the purpose of describing the embodiments of the present application.
[0021] The embodiments of the present application provide a power distribution network fault location system under interconnection perception, as shown in Figure 1 The system comprises:
[0022] The fault type identification module 10 is used to extract the initial operation perception waveform data set of the target power distribution network according to the interconnection perception network, to identify the predictive fault type of the initial operation perception waveform data set, and to determine the predictive fault type label.
[0023] Further, the specific configuration of the fault type identification module 10 further comprises collecting the operation perception waveform sample set of the target power distribution network under the known fault type label sample; performing pre-sequence perception waveform sample set extraction on the operation perception waveform sample set, obtaining a pre-sequence perception waveform sample set, and identifying a predictive waveform feature vector sample of the pre-sequence perception waveform sample set; training a predictive fault type identifier according to the predictive waveform feature vector sample and the known fault type label sample, using the predictive fault type identifier to identify the predictive fault type of the initial operation perception waveform data set, and determining the predictive fault type label.
[0024] Preferably, the fault type identification module is a data processing classification unit for power distribution network fault analysis, which judges the possible fault type by feature extraction and pattern recognition on the power grid operation waveform data. Specifically, by deploying an interconnected perception network in the target power distribution network, voltage, current and other waveform data in the power grid are continuously collected in a high-frequency sampling manner to obtain an initial operation perception waveform data set, which contains waveform sequences of multiple time points and multiple monitoring points. The interconnected perception network may include intelligent sensors, synchronous phasor measurement devices, fault recording devices, etc.
[0025] Preferably, the predictive fault type identification is performed on the initial operation perception waveform data set. Specifically, waveform samples under the known fault type label samples such as single-phase ground fault type, phase-to-phase short circuit fault type, two-phase ground short circuit fault type, three-phase short circuit fault and line breakage fault in the historical data of the target power distribution network are collected to obtain an operation perception waveform sample set. The pre-sequence perception waveform sample set is extracted from the operation perception waveform sample set, i.e. the waveform segment before the actual occurrence of the fault or the beginning instant is intercepted from the operation perception waveform sample set to capture the early signs of the fault, determine the pre-sequence perception waveform sample set, and identify the predictive waveform feature vector sample of the pre-sequence perception waveform sample set, including the amplitude mutation rate, harmonic distortion rate, phase angle change, waveform entropy, energy value of a specific frequency band, etc.
[0026] Preferably, a deep learning model such as a convolutional neural network (CNN) and a long short-term memory (LSTM) is used to construct an identification model, the predictive waveform feature vector sample is used as the input, and the known fault type label sample is used as the target output to train the identification model, thereby learning the mapping relationship between the predictive waveform feature vector and the fault type, finally obtaining a predictive fault type identifier, and then using the predictive fault type identifier to identify the predictive fault type of the initial operation perception waveform data set. Specifically, the same preprocessing and feature extraction are performed on the initial operation perception waveform data set to obtain real-time waveform feature vectors, which are then input into the predictive fault type identifier for identification, outputting the most possible fault type prediction, and finally determining the predictive fault type label.
[0027] Further, the specific configuration of the fault type identification module 10 further comprises that the known fault type label sample at least includes a single-phase ground fault type, an inter-phase short-circuit fault type, a two-phase ground short-circuit fault type, a three-phase short-circuit fault, and a broken line fault; and the operation perception waveform sample set includes operation perception waveform samples under at least three different fault levels.
[0028] Preferably, the known fault type label sample at least includes a single-phase ground fault type, an inter-phase short-circuit fault type, a two-phase ground short-circuit fault type, a three-phase short-circuit fault, and a broken line fault, wherein the single-phase ground fault refers to a short circuit between any one of the A, B, and C phases and the ground in the power grid, which is the fault type with the highest occurrence rate; the inter-phase short-circuit fault refers to a direct short circuit between two-phase conductors of the A-B phase, the B-C phase, or the A-C phase; the two-phase ground short-circuit fault refers to a short circuit formed by grounding of two-phase conductors; the three-phase short-circuit fault refers to a short circuit of the A, B, and C three-phase conductors, which is the most serious fault type but has a low occurrence rate; and the broken line fault refers to a broken line caused by external force or aging, resulting in non-full-phase operation. Different fault types exhibit completely different characteristics in waveforms, for example, a ground fault generates zero sequence current, while an inter-phase short-circuit fault does not. For the same fault type, the operation perception waveform sample set includes operation perception waveform samples under at least three different fault levels, wherein the fault level refers to the size of the fault transition resistance, and the transition resistance is the resistance value of the fault point, which greatly affects the amplitude and waveform characteristics of the fault current. The transition resistance of a low-resistance fault is very small, which may be close to 0 ohm, the fault current is very large, and the waveform distortion is very obvious; the transition resistance of a medium-resistance fault is several tens to several hundred ohms, the fault current is moderate, and the waveform characteristics are not as obvious as those of a low-resistance fault; and the transition resistance of a high-resistance fault is very large, the fault current is very small, the characteristics are extremely weak, and it is similar to normal load current, which is difficult to detect.
[0029] The mapping library construction module 20 is configured to construct a fault type-visibility window mapping library, including a mapping relationship between a fault type label and a visibility time window, and the visibility time window is a time window in which a feature visibility index of the interconnected perception network for interconnected perception data is greater than a preset feature visibility index threshold.
[0030] Preferably, the mapping library construction module is used to establish a mapping relationship between fault types and time periods of occurrence of key features, i.e., a mapping relationship between fault type labels and explicit time windows, and thus obtain a fault type-explicit window mapping library, wherein the explicit time window is a time window in which a feature explicitness indicator of interconnected sensing network interconnected sensing data is greater than a preset feature explicitness indicator threshold, i.e., a time period in which key features in the fault waveform data can most clearly and most significantly represent fault types, the feature explicitness indicator is used to quantify the degree of distinctness of a certain fault feature, for example, for a short-circuit fault, the feature explicitness indicator can be a current mutation rate or a total harmonic distortion rate, and the greater the value, the more obvious the feature; the preset feature explicitness indicator threshold is a pre-set threshold value for judging whether a feature is distinct, when the feature explicitness indicator exceeds the preset feature explicitness indicator threshold, the feature is considered to be explicit at this moment, and finally all time points at which the feature explicitness indicator exceeds the preset feature explicitness indicator threshold are connected to form the explicit time window.
[0031] Further, the specific configuration of the mapping library construction module 20 further includes extracting a center sensing waveform sample set from the running sensing waveform sample set, obtaining a center sensing waveform sample set, and identifying a center waveform feature vector sample of the center sensing waveform sample set; analyzing the center waveform feature vector sample to determine key waveform features corresponding to each known fault type label; defining an explicitness determination rule of the key waveform features, analyzing the center sensing waveform sample set according to the explicitness determination rule to obtain explicit time window samples of the key waveform features; fitting a mapping relationship between each known fault type label and the explicit time window samples to obtain a fault type-explicit window mapping library.
[0032] Preferably, the center sensing waveform sample set is extracted from the running sensing waveform sample set, wherein the center sensing waveform refers to a waveform in a stable development stage of a fault, i.e., a waveform that is most sufficient and stable in feature after the occurrence of a fault is extracted from a complete fault waveform sample, the center sensing waveform sample set is obtained, high-precision feature recognition extraction is performed on the center sensing waveform sample set, and a center waveform feature vector sample is constructed, which can include a current waveform mutation point, a zero-sequence current amplitude, a negative-sequence current amplitude, a fault-phase current amplitude, etc. Then, the center waveform feature vector sample is analyzed by using a decision tree and a random forest feature importance evaluation to determine which of the multiple features are most representative and most critical for distinguishing different fault types, and thus key waveform features most indicative of each fault type are determined, for example, for a ground fault, the key waveform feature is the amplitude of the zero-sequence current; for an inter-phase short-circuit fault, the key waveform feature is the amplitude of the negative-sequence current or the amplitude of the fault-phase current; for an arc fault, the key waveform feature is the number of current waveform mutations or the energy of a specific frequency band; and for a broken line fault, the key waveform feature is the negative-sequence current phase.
[0033] Preferably, the explicit judgment rule defining the key waveform feature is that if the feature explicitness indicator of the key feature is greater than the preset feature explicitness indicator threshold value, it is considered that the key waveform feature is explicit at this moment, then the central perception waveform sample set is rescanned and analyzed according to the explicit judgment rule, and all time periods during which the feature explicitness indicator of each key waveform feature is continuously greater than the preset feature explicitness indicator threshold value are identified and determined as the explicit time window sample of the key waveform feature; finally, the mapping relationship between each known fault type label and the explicit time window sample is fitted, specifically, for the same fault type, all explicit time window samples thereof are analyzed, the intersection of all the explicit time window samples is taken to fit and determine the general explicit time window representing the fault type, and then the mapping relationship between the known fault type label and the explicit time window sample is fitted to form a fault type-explicit window mapping library.
[0034] Further, the specific configuration of the mapping library construction module 20 further includes that the explicit judgment rule includes an explicit judgment triple, the explicit judgment triple includes a key waveform feature indicator, a feature explicitness indicator, and an explicit judgment threshold value; the feature explicitness indicator of the current key waveform feature is identified according to the explicit judgment rule, and the corresponding explicit judgment threshold value is extracted according to the feature explicitness indicator; the central perception waveform sample set is analyzed based on the feature explicitness indicator, the explicit perception waveform sample satisfying the explicit judgment threshold value is obtained through binary judgment, and the time window corresponding to the explicit perception waveform sample is marked as the explicit time window sample of the current key waveform feature.
[0035] Preferably, the explicit judgment rule includes an explicit judgment triple, the explicit judgment triple includes a key waveform feature indicator, such as a zero sequence current amplitude, an A-phase voltage effective value, a 5th harmonic content, a current waveform kurtosis, etc.; a feature explicitness indicator, such as an instantaneous value of the zero sequence current amplitude, a total harmonic distortion rate of the harmonic content; and an explicit judgment threshold value, which is a preset numerical threshold value, used for comparison with the feature explicitness indicator.
[0036] Preferably, the explicit judgment rule is that when the feature explicitness indicator is greater than the preset feature explicitness indicator threshold value, it is considered that the key feature is explicit at this moment, the feature explicitness indicator of the current key waveform feature is identified and determined according to the explicit judgment rule, and the corresponding explicit judgment threshold value is set according to the feature explicitness indicator, for example, the feature explicitness indicator of the zero sequence current amplitude is the ratio of the current instantaneous amplitude or effective value to the normal amplitude or rated value before the fault, and the explicit judgment threshold value is set to 1.2, indicating that it is considered to be explicit when it is 20% higher than the normal amplitude; the feature explicitness indicator of the signal-to-noise ratio is the ratio of the energy of the fault feature to the energy of the background noise, and the explicit judgment threshold value is set to 10; the feature explicitness indicator of the high-frequency transient is the wavelet coefficient energy of a specific frequency band after wavelet transform of the waveform, and the explicit judgment threshold value is determined by statistics of a large number of normal and fault samples.
[0037] Preferably, the feature saliency index analysis center perception waveform sample set is analyzed by binary judgment to determine point-by-point scanning of the entire center perception waveform sample, that is, traversing the feature saliency index of each time point of the center perception waveform, and then comparing it with the corresponding saliency judgment threshold to output the binary result. If the feature saliency index is greater than the saliency judgment threshold, mark the time point as 1, otherwise mark the time point as 0. Then the corresponding perception waveform of the time point marked as 1 is taken as the salient perception waveform sample, and the salient perception waveform sample corresponding time window is marked as the salient time window sample of the current key waveform feature.
[0038] Further, the specific configuration of the mapping library construction module 20 further comprises: obtaining the first key waveform feature and the second key waveform feature of the current known fault type label; obtaining the first salient time window and the second salient time window according to the first key waveform feature and the second key waveform feature; and taking the time window intersection of the first salient time window and the second salient time window as the label salient time window of the current known fault type mapping.
[0039] Preferably, the current known fault type label is analyzed to determine at least two representative key waveform features as the first key waveform feature and the second key waveform feature. For example, for single-phase ground fault, the first key waveform feature is the significant increase of zero sequence current amplitude, and the second key waveform feature is the significant decrease of single-phase voltage amplitude. According to the saliency judgment rule, all time periods in which the first key waveform feature and the second key waveform feature are significantly present are determined respectively to obtain the first salient time window and the second salient time window. Then, the first salient time window and the second salient time window are intersected on the time axis to determine the time period that satisfies all key waveform features as the label salient time window of the fault type mapping, thereby minimizing the data processing amount and improving the reliability.
[0040] The salient data extraction module 30 is configured to map the predictive fault type label by using the fault type-salient window mapping library, determine the mapping salient time window for the interconnected perception network, and extract the salient running perception waveform data set under the mapping salient time window.
[0041] Preferably, the explicit data extraction module is an intelligent data gateway connected with fault identification and accurate positioning, used to provide high-value data on demand, specifically, a predictive fault type label is taken as a Key to match and map in a fault type-explicit window mapping library, to obtain a mapped explicit time window for an interconnected sensing network, and then the initial operation sensing waveform data set is extracted using the determined mapped explicit time window to accurately intercept the waveform data corresponding to the time window, and finally determine the explicit operation sensing waveform data set, i.e. multiple waveform data segments with greatly shortened length and significantly improved signal-to-noise ratio, thereby ensuring the rapidity and accuracy of the fault positioning process.
[0042] The fault positioning result acquisition module 40 is configured to perform interconnected sensing amplification processing on the explicit operation sensing waveform data set to obtain an amplified explicit operation sensing waveform data set, and input the amplified explicit operation sensing waveform data set into a next-stage fault positioning unit for analysis to obtain a fault positioning result.
[0043] Preferably, the fault positioning result acquisition module is configured to perform data enhancement on the extracted high-quality data and schedule fault positioning, specifically, the explicit operation sensing waveform data set is subjected to interconnected sensing amplification processing, including targeted data preprocessing and feature enhancement, i.e. the waveform data from different monitoring points within the same time window are subjected to spatio-temporal alignment and fusion, and a wavelet transform or other digital signal processing algorithm is used to transform the waveform, highlight the transient or steady-state features related to fault positioning, and suppress irrelevant noise and load fluctuations, thereby obtaining the amplified explicit operation sensing waveform data set; and then input the amplified explicit operation sensing waveform data set into a next-stage fault positioning unit for analysis, wherein the next-stage fault positioning unit can be a traveling wave positioning unit, an impedance method positioning unit, an intelligent algorithm positioning unit, or the like, to obtain accurate fault point position information as the fault positioning result and improve the fault identification and positioning accuracy of the distribution network.
[0044] Further, the fault positioning result acquisition module 40 is further configured to read the fault positioning algorithm type of the next-stage fault positioning unit, determine a waveform input template for fault positioning according to the fault positioning algorithm type, and perform interconnected sensing amplification processing on the explicit operation sensing waveform data set according to the waveform input template to obtain the amplified explicit operation sensing waveform data set.
[0045] Preferably, the type of fault location algorithm adopted by the next-stage fault location unit is read through the query interface, such as a traveling wave location algorithm, an impedance method location algorithm, or an artificial intelligence-based location algorithm, and a waveform input template for fault location is determined according to the type of fault location algorithm. Specifically, for the traveling wave location algorithm, the waveform input template may require a very high data sampling rate, a very short data length, and data content of high-frequency transient components of three-phase current / voltage; for the impedance method location algorithm, the waveform input template may require a power frequency sampling rate, a steady-state data length of several cycles after the fault, and data content of fundamental effective values and phase angles of voltage and current. Finally, the explicit operation perception waveform data set is interconnected and perception amplified according to the waveform input template, that is, an accurate signal processing operation is performed, which may include resampling, filtering processing, data cropping and alignment, feature calculation and conversion, etc., and finally an amplified explicit operation perception waveform data set is output, thereby fundamentally guaranteeing the accuracy, speed and reliability of fault location.
[0046] In the foregoing, reference is made to Figure 1 The power distribution network fault location system under interconnected perception according to the embodiments of the application is described in detail. Next, the power distribution network fault location method under interconnected perception according to the embodiments of the application will be described with reference to Figure 2 The power distribution network fault location method under interconnected perception according to the embodiments of the application will be described with reference to
[0047] The power distribution network fault location method under interconnected perception, as shown in Figure 2 includes: extracting an initial operation perception waveform data set of a target power distribution network according to an interconnected perception network, performing predictive fault type identification on the initial operation perception waveform data set to determine a predictive fault type label; constructing a fault type-explicit window mapping library including a mapping relationship between a fault type label and an explicit time window, the explicit time window being a time window in which a feature explicitness index of interconnected perception data of the interconnected perception network is greater than a preset feature explicitness index threshold; mapping the predictive fault type label through the fault type-explicit window mapping library to determine a mapped explicit time window for the interconnected perception network and extract an explicit operation perception waveform data set under the mapped explicit time window; performing interconnected perception amplification processing on the explicit operation perception waveform data set to obtain an amplified explicit operation perception waveform data set, and inputting the amplified explicit operation perception waveform data set into a next-stage fault location unit for analysis to obtain a fault location result.
[0048] In a possible implementation, the power distribution network fault location method under interconnected perception further includes: collecting a running perception waveform sample set of the target power distribution network under a known fault type label sample; performing presequence perception waveform sample set extraction on the running perception waveform sample set, obtaining a presequence perception waveform sample set, and identifying a predictability waveform feature vector sample of the presequence perception waveform sample set; performing deep training on the predictability waveform feature vector sample and the known fault type label sample to obtain a predictability fault type identifier, identifying a predictability fault type of the initial running perception waveform data set by using the predictability fault type identifier, and determining a predictability fault type label.
[0049] In a possible implementation, the power distribution network fault location method under interconnected perception further includes: the known fault type label sample at least includes a single-phase grounding fault type, an inter-phase short-circuit fault type, a two-phase grounding short-circuit fault type, a three-phase short-circuit fault, and a broken line fault; and the running perception waveform sample set includes running perception waveform samples under at least three different fault levels.
[0050] In a possible implementation, the power distribution network fault location method under interconnected perception further includes: performing center perception waveform sample set extraction on the running perception waveform sample set, obtaining a center perception waveform sample set, and identifying a center waveform feature vector sample of the center perception waveform sample set; analyzing the center waveform feature vector sample to determine a key waveform feature corresponding to each known fault type label; defining a dominant decision rule of the key waveform feature, analyzing the center perception waveform sample set according to the dominant decision rule to obtain a dominant time window sample of the key waveform feature; and fitting a mapping relationship between each known fault type label and the dominant time window sample to obtain a fault type-dominant window mapping library.
[0051] In a possible implementation, the power distribution network fault location method under interconnected perception further includes: obtaining a first key waveform feature and a second key waveform feature of a current known fault type label; obtaining a first dominant time window and a second dominant time window according to the first key waveform feature and the second key waveform feature; and taking a time window intersection of the first dominant time window and the second dominant time window as a label dominant time window of a current known fault type mapping.
[0052] In a possible implementation, the power distribution network fault locating method under interconnection perception further includes that: the explicit decision rule includes an explicit decision triad, the explicit decision triad includes a key waveform feature index, a feature explicitness index, and an explicit decision threshold; a feature explicitness index of a current key waveform feature is identified according to the explicit decision rule, and a corresponding explicit decision threshold is extracted according to the feature explicitness index; the central perception waveform sample set is analyzed based on the feature explicitness index, an explicit perception waveform sample meeting the explicit decision threshold is obtained through binary decision, and a time window corresponding to the explicit perception waveform sample is marked as an explicit time window sample of the current key waveform feature.
[0053] In a possible implementation, the power distribution network fault locating method under interconnection perception further includes: reading a fault locating algorithm type of the next-level fault locating unit, determining a waveform input template for fault locating according to the fault locating algorithm type; and performing interconnection perception amplification processing on the explicit operation perception waveform data set according to the waveform input template, to obtain an amplified explicit operation perception waveform data set.
[0054] The power distribution network fault locating system under interconnection perception provided in the embodiments of the present application can execute the power distribution network fault locating method under interconnection perception provided in any of the embodiments of the present application, and has the corresponding function modules and beneficial effects of the execution method.
[0055] Although various references are made to certain modules in the system according to the embodiments of the present application, however, any number of different modules can be used and run on the user terminal and / or server, and each unit and module included is only divided according to the functional logic, but is not limited to the above division, as long as the corresponding functions can be realized; in addition, the specific names of each functional unit are only for easy mutual differentiation, and do not limit the protection scope of the present application.
[0056] The above specific embodiments do not constitute a limitation on the protection scope of the present application. Those skilled in the art should understand that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent replacement and improvement made within the spirit and principle of the present application should be included in the protection scope of the present application.
Claims
1. A power distribution network fault location system under inter-connected perception, characterized in that, The system comprises: a fault type identification module, configured to extract an initial operation perception waveform data set of a target power distribution network according to an interconnected perception network, perform predictive fault type identification on the initial operation perception waveform data set, and determine a predictive fault type label; a mapping library construction module, configured to construct a fault type-explicit window mapping library, including a mapping relationship between a fault type label and an explicit time window, the explicit time window being a time window in which a feature explicitness index of interconnected perception data of the interconnected perception network is greater than a preset feature explicitness index threshold; an explicit data extraction module, configured to map the predictive fault type label through the fault type-explicit window mapping library, determine a mapping explicit time window for the interconnected perception network, and extract an explicit operation perception waveform data set in the mapping explicit time window; a fault location result acquisition module, configured to perform interconnected perception amplification processing on the explicit operation perception waveform data set, acquire an amplified explicit operation perception waveform data set, input the amplified explicit operation perception waveform data set into a next-level fault location unit for analysis, and acquire a fault location result.
2. The power distribution network fault location system under interlinked awareness of claim 1, wherein, The predictive fault type identification on the initial operation perception waveform data set and the determination of the predictive fault type label include: collecting an operation perception waveform sample set of the target power distribution network under a known fault type label sample; extracting a presequence perception waveform sample set from the operation perception waveform sample set, acquiring a presequence perception waveform sample set, and identifying a predictive waveform feature vector sample of the presequence perception waveform sample set; training a predictive fault type identifier according to the predictive waveform feature vector sample and the known fault type label sample, and identifying a predictive fault type of the initial operation perception waveform data set by using the predictive fault type identifier to determine a predictive fault type label.
3. The power distribution system fault location system under interlinked awareness as claimed in claim 2 wherein, The known fault type label sample at least includes a single-phase ground fault type, an inter-phase short-circuit fault type, a two-phase ground short-circuit fault type, a three-phase short-circuit fault, and a broken line fault; The operation perception waveform sample set includes operation perception waveform samples in at least three different fault levels.
4. The power distribution system fault location system under interconnected awareness of claim 2, wherein, The construction of the fault type-explicit window mapping library includes: extracting a center perception waveform sample set from the operation perception waveform sample set, acquiring a center perception waveform sample set, and identifying a center waveform feature vector sample of the center perception waveform sample set; analyzing the center waveform feature vector sample to determine a key waveform feature corresponding to each known fault type label; defining an explicitness determination rule of the key waveform feature, analyzing the center perception waveform sample set according to the explicitness determination rule to obtain an explicit time window sample of the key waveform feature; fitting a mapping relationship between each known fault type label and the explicit time window sample to obtain a fault type-explicit window mapping library.
5. The power distribution system fault location system under interconnected awareness of claim 4, wherein, The fitting of the mapping relationship between each known fault type label and the explicit time window sample further includes: acquiring a first key waveform feature and a second key waveform feature of a current known fault type label; obtaining a first dominant time window and a second dominant time window according to the first key waveform feature and the second key waveform feature; taking a time window intersection of the first dominant time window and the second dominant time window as a label dominant time window of a current known fault type mapping.
6. The power distribution system fault location system under interconnected awareness of claim 4, wherein, analyzing the central perception waveform sample set according to the dominant determination rule to obtain a dominant time window sample of the key waveform feature, comprising: wherein the dominant determination rule comprises a dominant determination triple, and the dominant determination triple comprises a key waveform feature index, a feature dominant index, and a dominant determination threshold value; identifying a feature dominant index of a current key waveform feature according to the dominant determination rule, and extracting a corresponding dominant determination threshold value according to the feature dominant index; analyzing the central perception waveform sample set based on the feature dominant index, obtaining a dominant perception waveform sample satisfying the dominant determination threshold value through binary determination, and marking a time window corresponding to the dominant perception waveform sample as a dominant time window sample of the current key waveform feature.
7. The power distribution system fault location system under interconnected awareness of claim 1, wherein, interconnecting and amplifying the dominant running perception waveform data set to obtain an amplified dominant running perception waveform data set, comprising: reading a fault location algorithm type of the next-stage fault location unit, and determining a waveform input template used for fault location according to the fault location algorithm type; interconnecting and amplifying the dominant running perception waveform data set according to the waveform input template to obtain an amplified dominant running perception waveform data set.
8. A method for fault location in a distribution network under interconnection awareness, characterized in that, The method is applied to the power distribution network fault location system under interconnection perception of any one of claims 1-7, and the method comprises: extracting an initial running perception waveform data set of a target power distribution network according to an interconnection perception network, performing predictive fault type identification on the initial running perception waveform data set, and determining a predictive fault type label; constructing a fault type-dominant window mapping library comprising a mapping relationship between a fault type label and a dominant time window, wherein the dominant time window is a time window in which a feature dominant index of interconnection perception data of the interconnection perception network is greater than a preset feature dominant index threshold value; mapping the predictive fault type label through the fault type-dominant window mapping library, determining a mapping dominant time window for the interconnection perception network, and extracting a dominant running perception waveform data set under the mapping dominant time window; interconnecting and amplifying the dominant running perception waveform data set to obtain an amplified dominant running perception waveform data set, inputting the amplified dominant running perception waveform data set into a next-stage fault location unit for analysis, and obtaining a fault location result.