Defect identification method and device for power distribution network, computer device and program product
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- GUANGZHOU POWER SUPPLY BUREAU GUANGDONG POWER GRID CO LTD
- Filing Date
- 2026-04-13
- Publication Date
- 2026-07-21
Smart Images
Figure CN122432624A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of defect identification technology, and in particular to a defect identification method, apparatus, computer equipment, computer-readable storage medium, and computer program product for power distribution networks. Background Technology
[0002] The power distribution network is the core link connecting power sources and users in the power system, and its operational stability directly affects the quality of power supply. With the expansion of the distribution network, the increase in equipment types, and the complexity of the operating environment, various defects such as line grounding faults, switch failures, communication link interruptions, and equipment overheating occur frequently.
[0003] In related technologies, existing methods for identifying distribution network defects mainly rely on manual inspections or monitoring and analysis from a single data source. Manual inspections are inefficient and costly, making it difficult to achieve full coverage and detect hidden defects in a timely manner. Information collected from a single data source is incomplete, failing to integrate data from multiple links such as the distribution master station, distribution terminals, and communication links, resulting in data quality issues such as anomalies, redundancy, and inconsistent formats. Furthermore, existing methods often focus on comparing single features, failing to fully explore the temporal and spatial correlations between data, leading to low accuracy and intelligence in defect identification and ultimately, inaccurate identification results. Summary of the Invention
[0004] Therefore, it is necessary to provide a method, device, computer equipment, computer-readable storage medium, and computer program product for identifying distribution network defects that can improve the accuracy of distribution network defect identification, in response to the above-mentioned technical problems.
[0005] Firstly, this application provides a method for identifying defects in a power distribution network, including:
[0006] Collect full-domain information data of the power distribution network, and perform data cleaning and standardization on the full-domain information data to obtain a standardized dataset;
[0007] Extract the temporal feature vector and spatial feature vector from the standardized dataset, and fuse the temporal feature vector and the spatial feature vector to obtain the target feature vector;
[0008] The target feature vector is compared with the sample feature vector of each defect type in the preset defect library to obtain a preliminary similarity value.
[0009] Based on the standardized dataset and the network topology of the distribution network, the temporal correlation features and spatial correlation features of the distribution network are extracted.
[0010] Based on the preliminary similarity value, the temporal correlation feature, and the spatial correlation feature, the defect type of the distribution network is determined.
[0011] In one embodiment, the step of performing data cleaning and standardization on the global information data to obtain a standardized dataset includes:
[0012] Remove outliers from the global information data, and use the device ID, timestamp, and data type in the power distribution network as a joint key to delete duplicate data, thus obtaining the deleted global information data.
[0013] The missing data in the deleted global information data is filled in to obtain the filled global information data.
[0014] The data format of the filled global information data is normalized, and a standardized dataset of the power distribution network is constructed based on the association mapping relationship between the device ID, timestamp and data type.
[0015] In one embodiment, the step of extracting the temporal correlation features and spatial correlation features of the distribution network based on the standardized dataset and the network topology of the distribution network includes:
[0016] Extract the time dependencies of the standardized data in the standardized dataset in the time dimension, and generate the time-series correlation features of the distribution network based on the time dependencies;
[0017] Based on the standardized dataset and the network topology, the spatial dependencies of the power distribution network are extracted, and the spatial correlation features of the power distribution network are generated based on the spatial dependencies.
[0018] In one embodiment, determining the defect type of the distribution network based on the preliminary similarity value, the temporal correlation feature, and the spatial correlation feature includes:
[0019] Calculate the temporal correlation similarity value between the temporal correlation feature and the temporal correlation feature of each defect type in the preset defect library, and the spatial correlation similarity value between the spatial correlation feature and the spatial correlation feature of each defect type in the preset defect library;
[0020] Preset weights are assigned to the preliminary similarity value, the temporal correlation feature similarity value, and the spatial correlation feature similarity value, and a weighted sum is performed to obtain the final similarity of each defect type in the distribution network;
[0021] The defect types with a final similarity greater than a preset defect threshold are taken as the defect types of the power distribution network.
[0022] In one embodiment, the defect types include line grounding faults, line short-circuit faults, switch failure to operate defects, switch malfunction defects, communication link interruption defects, equipment overheating defects, transformer fault defects, and load overload defects.
[0023] In one embodiment, the global information data includes power distribution master station data, power distribution terminal data, and communication link data;
[0024] The data from the distribution master station includes the operating parameters of the distribution network, equipment control commands, dispatch commands, and preliminary fault assessment records; the operating parameters include voltage, current, power, and frequency.
[0025] The power distribution terminal data includes switchgear status, transformer monitoring data, equipment temperature data, partial discharge signals, and terminal alarm information; the switchgear status includes opening and closing status and fault trip records.
[0026] The communication link data includes communication latency, data packet loss rate, link connectivity status, signal strength, and transmission bandwidth utilization.
[0027] In one embodiment, the process of constructing the preset defect library includes:
[0028] Historical global information data for each defect type was collected, and the data was cleaned and standardized to obtain a historical standard dataset.
[0029] Extract the historical time-series feature vector and historical spatial feature vector from the historical standardized dataset, and fuse the historical time-series feature vector and the historical spatial feature vector to obtain the corresponding initial sample feature vector for each defect type;
[0030] For each defect type, the feature vectors of multiple initial samples are statistically analyzed according to feature dimensions, and the deviation range of each feature dimension is calculated.
[0031] Remove abnormal initial sample feature vectors that exceed the deviation range; perform statistical analysis on the remaining initial sample feature vectors for each defect type according to the feature dimension to obtain the feature mean of each feature dimension under each defect type, and construct the corresponding sample feature vector for each defect type.
[0032] A pre-defined defect library is constructed based on the sample feature vector corresponding to each defect type.
[0033] Secondly, this application also provides a defect identification device for a power distribution network, comprising:
[0034] The processing module is used to collect full-domain information data of the power distribution network, perform data cleaning and standardization processing on the full-domain information data, and obtain a standardized dataset;
[0035] The fusion module is used to extract the temporal feature vector and spatial feature vector from the standardized dataset, and fuse the temporal feature vector and the spatial feature vector to obtain the target feature vector;
[0036] The comparison module is used to compare the target feature vector with the sample feature vector of each defect type in the preset defect library to obtain a preliminary similarity value.
[0037] The extraction module is used to extract the temporal correlation features and spatial correlation features of the power distribution network based on the standardized dataset and the network topology of the power distribution network.
[0038] The determination module is used to determine the defect type of the distribution network based on the preliminary similarity value, the temporal correlation feature, and the spatial correlation feature.
[0039] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:
[0040] Collect full-domain information data of the power distribution network, and perform data cleaning and standardization on the full-domain information data to obtain a standardized dataset;
[0041] Extract the temporal feature vector and spatial feature vector from the standardized dataset, and fuse the temporal feature vector and the spatial feature vector to obtain the target feature vector;
[0042] The target feature vector is compared with the sample feature vector of each defect type in the preset defect library to obtain a preliminary similarity value.
[0043] Based on the standardized dataset and the network topology of the distribution network, the temporal correlation features and spatial correlation features of the distribution network are extracted.
[0044] Based on the preliminary similarity value, the temporal correlation feature, and the spatial correlation feature, the defect type of the distribution network is determined.
[0045] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:
[0046] Collect full-domain information data of the power distribution network, and perform data cleaning and standardization on the full-domain information data to obtain a standardized dataset;
[0047] Extract the temporal feature vector and spatial feature vector from the standardized dataset, and fuse the temporal feature vector and the spatial feature vector to obtain the target feature vector;
[0048] The target feature vector is compared with the sample feature vector of each defect type in the preset defect library to obtain a preliminary similarity value.
[0049] Based on the standardized dataset and the network topology of the distribution network, the temporal correlation features and spatial correlation features of the distribution network are extracted.
[0050] Based on the preliminary similarity value, the temporal correlation feature, and the spatial correlation feature, the defect type of the distribution network is determined.
[0051] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:
[0052] Collect full-domain information data of the power distribution network, and perform data cleaning and standardization on the full-domain information data to obtain a standardized dataset;
[0053] Extract the temporal feature vector and spatial feature vector from the standardized dataset, and fuse the temporal feature vector and the spatial feature vector to obtain the target feature vector;
[0054] The target feature vector is compared with the sample feature vector of each defect type in the preset defect library to obtain a preliminary similarity value.
[0055] Based on the standardized dataset and the network topology of the distribution network, the temporal correlation features and spatial correlation features of the distribution network are extracted.
[0056] Based on the preliminary similarity value, the temporal correlation feature, and the spatial correlation feature, the defect type of the distribution network is determined.
[0057] The aforementioned defect identification method, device, computer equipment, computer-readable storage medium, and computer program product for distribution networks firstly collect comprehensive information data of the distribution network, clean and standardize the comprehensive information data to obtain a standardized dataset; extract temporal and spatial feature vectors from the standardized dataset, and fuse the temporal and spatial feature vectors to obtain a target feature vector; compare the target feature vector with the sample feature vectors of each defect type in a preset defect database to obtain a preliminary similarity value; based on the standardized dataset and the network topology of the distribution network, extract the temporal and spatial correlation features of the distribution network; and based on the preliminary similarity value, temporal correlation features, and spatial correlation features, determine the defect type of the distribution network. Thus, by constructing a comprehensive data collection system covering the distribution master station, distribution terminals, and communication links, and adopting a combined real-time and periodic acquisition method, comprehensive coverage and efficient acquisition of distribution network operation information were achieved. Based on this, a combined data cleaning algorithm and standardized integration model were used to effectively remove abnormal and redundant data, fill in missing values, and unify the format of multi-source data, significantly improving data quality and usability. Furthermore, by constructing an expert database containing common defects and spatiotemporal feature vectors, and combining cosine similarity algorithm for preliminary multi-dimensional feature comparison, and innovatively employing temporal trajectory analysis and spatial trajectory analysis, the evolutionary patterns of data in the time dimension and the topological relationships in the spatial dimension were deeply explored. Finally, through weighted fusion and threshold determination, accurate and automatic defect identification was achieved. Attached Figure Description
[0058] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.
[0059] Figure 1 This is an application environment diagram of a defect identification method for a power distribution network in one embodiment;
[0060] Figure 2 This is a flowchart illustrating a defect identification method for a power distribution network in one embodiment;
[0061] Figure 3 This is a flowchart illustrating a defect identification method for a power distribution network in another embodiment;
[0062] Figure 4 This is a structural block diagram of a defect identification device for a power distribution network in one embodiment;
[0063] Figure 5This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation
[0064] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.
[0065] It should be noted that the terms "first," "second," etc., used in this application can be used to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish the first element from the second element. The terms "comprising" and "having," and any variations thereof, used in this application, are intended to cover non-exclusive inclusion. The term "multiple" used in this application refers to two or more. The term "and / or" used in this application refers to one of the embodiments, or any combination of multiple embodiments.
[0066] The defect identification method for power distribution networks provided in this application can be applied to, for example... Figure 1 In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated onto server 104 or located on the cloud or other network servers. Terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, drones, low-altitude aircraft, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, projection devices, etc. Portable wearable devices can include smartwatches, smart bracelets, head-mounted devices, etc. Head-mounted devices can be virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, etc. Server 104 can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services.
[0067] In one exemplary embodiment, such as Figure 2 As shown, a defect identification method for a power distribution network is provided, which can be applied to... Figure 1 Taking terminal 102 as an example, the explanation includes the following steps 202 to 210. Wherein:
[0068] Step 202: Collect full-domain information data of the power distribution network, perform data cleaning and standardization on the full-domain information data to obtain a standardized dataset.
[0069] For example, a comprehensive data collection system is constructed to collect comprehensive information data of the distribution network, and the comprehensive information data is cleaned and standardized to obtain a standardized dataset of the distribution network.
[0070] Step 204: Extract the temporal feature vector and spatial feature vector from the standardized dataset, and fuse the temporal feature vector and spatial feature vector to obtain the target feature vector.
[0071] Optionally, the temporal feature vector and spatial feature vector of each monitored object are extracted from the standardized dataset, and the temporal feature vector and spatial feature vector are fused to obtain the target feature vector.
[0072] The method for fusing temporal feature vectors and spatial feature vectors is the same as the method for fusing feature vectors of each defect type sample in the preset defect library, and this application embodiment does not limit this.
[0073] Step 206: Compare the target feature vector with the sample feature vector of each defect type in the preset defect library to obtain a preliminary similarity value.
[0074] For example, the cosine similarity between the target feature vector and the sample feature vector of each defect type in the preset defect library is calculated. The calculation formula is shown in formula (1) to obtain the preliminary similarity value between the target feature vector and each defect type.
[0075] (1)
[0076] in, The extracted feature vector; The feature vectors in the expert database; The cosine of the angle between the vectors is the preliminary similarity value, which reflects the directional consistency of the two vectors. The closer the value is to 1, the higher the feature matching degree.
[0077] Step 208: Based on the standardized dataset and the network topology of the distribution network, extract the temporal correlation features and spatial correlation features of the distribution network.
[0078] Optionally, based on standardized datasets and the network topology of the distribution network, trajectory analysis algorithms are used to mine the temporal and spatial correlation features of the standardized data.
[0079] Step 210: Based on the preliminary similarity value, temporal correlation characteristics, and spatial correlation characteristics, determine the defect type of the distribution network.
[0080] For example, temporal correlation features and spatial correlation features are incorporated into the preliminary similarity value to calculate the final similarity value. When the final similarity value exceeds a preset threshold, the corresponding defect type is determined.
[0081] In the aforementioned method for identifying defects in distribution networks, the following steps are taken: First, comprehensive information data of the distribution network is collected. This data is then cleaned and standardized to obtain a standardized dataset. Second, temporal and spatial feature vectors are extracted from the standardized dataset and fused to obtain a target feature vector. Third, the target feature vector is compared with the sample feature vectors of each defect type in a pre-defined defect database to obtain a preliminary similarity value. Finally, based on the standardized dataset and the network topology of the distribution network, temporal and spatial correlation features are extracted. Based on the preliminary similarity value, temporal correlation features, and spatial correlation features, the defect type of the distribution network is determined. Thus, by constructing a comprehensive data collection system covering the distribution master station, distribution terminals, and communication links, and adopting a combined real-time and periodic acquisition method, comprehensive coverage and efficient acquisition of distribution network operation information were achieved. Based on this, a combined data cleaning algorithm and standardized integration model were used to effectively remove abnormal and redundant data, fill in missing values, and unify the format of multi-source data, significantly improving data quality and usability. Furthermore, by constructing an expert database containing common defects and spatiotemporal feature vectors, and combining cosine similarity algorithm for preliminary multi-dimensional feature comparison, and innovatively employing temporal trajectory analysis and spatial trajectory analysis, the evolutionary patterns of data in the time dimension and the topological relationships in the spatial dimension were deeply explored. Finally, through weighted fusion and threshold determination, accurate and automatic defect identification was achieved.
[0082] In an exemplary embodiment, the global information data is cleaned and standardized to obtain a standardized dataset, including: removing outliers from the global information data and deleting duplicate data using the device ID, timestamp, and data type in the distribution network as joint keys to obtain the deleted global information data; filling in missing data in the deleted global information data to obtain the filled global information data; normalizing the data format of the filled global information data and constructing a standardized dataset of the distribution network based on the association mapping relationship between device ID, timestamp, and data type.
[0083] In practice, a combined data cleaning algorithm is used to remove abnormal and redundant data and fill in missing data, based on... Outlier removal algorithm based on principles: Let a certain data sequence be... Calculate the mean of the data and standard deviation It will exceed Data within the specified range is identified as outliers and removed.
[0084] Redundant data deletion algorithm based on duplication detection: using device ID, timestamp, and data type as joint keys, it identifies and deletes duplicate data records. For continuous data, linear interpolation is used, and the expression is shown in formula (2).
[0085] (2)
[0086] in, For timestamps of missing data, , The timestamps of valid data adjacent to the missing data.
[0087] In some embodiments, a standardized data integration model is constructed. First, the data formats of different data sources are unified, including timestamp format, data unit, and data encoding rules. Then, using the device unique identifier ID and timestamp as association keys, an association mapping relationship is established between different data sources to realize the association and fusion of power distribution master station data, power distribution terminal data, and communication link data, and output a standardized dataset.
[0088] In the above embodiments, by performing a series of data processing steps, the quality, consistency, and usability of the original data are significantly improved, providing a reliable data foundation for subsequent feature extraction, trajectory analysis, and defect identification, and ensuring the accuracy and robustness of defect identification from the source.
[0089] In an exemplary embodiment, based on a standardized dataset and the network topology of the distribution network, the temporal correlation features and spatial correlation features of the distribution network are extracted, including: extracting the temporal dependencies of standardized data in the standardized dataset in the time dimension, and generating the temporal correlation features of the distribution network based on the temporal dependencies; and extracting the spatial dependencies of the distribution network based on the standardized dataset and the network topology, and generating the spatial correlation features of the distribution network based on the spatial dependencies.
[0090] In practical implementation, LSTM (Long Short-Term Memory) network is used to perform time series correlation mining on the standardized data in the standardized dataset. For some pre-fault data in the standardized data, the temporal feature learning capability of the LSTM algorithm is used to perform in-depth mining on the continuous time series data of operating parameters such as voltage and current, and capture the time dependence of parameters over time, such as the trend of change of parameters and the fluctuation pattern (for example, the correlation between the trend of change of operating parameters before the fault and the fault type, and the correlation between periodic fluctuations and equipment aging defects).
[0091] In some embodiments, a fixed-length time window and step size are set according to the operating characteristics and defect identification requirements of the distribution network equipment. The standardized time series data mined by LSTM is truncated by sliding the window along the time axis to form several local time series data subsequences of equal length. Then, data statistics and feature calculations are performed on each subsequence to extract local feature indicators such as the mean, variance, extreme values, rate of change, and fluctuation amplitude of the operating parameters within the subsequence. At the same time, combined with the feature change trends between windows, key local time series patterns that can reflect the short-term operating status of the equipment are screened out. The key local time series patterns are integrated with the time dependency relationship to generate the time series correlation features of the distribution network.
[0092] In some embodiments, based on the network topology and standardized dataset of the power distribution network, a spatial correlation model is constructed using a graph neural network (GNN) to mine the spatial dependencies between different devices and different communication links (such as fault propagation correlation of adjacent lines, load coordination correlation of devices in the same area, and signal transmission correlation between communication links and power distribution terminals) and generate spatial correlation features.
[0093] In some embodiments, the spatial correlation model is based on the distribution network topology and constructed using the GNN algorithm. The specific process is as follows: First, the equipment and communication links in the distribution network are abstracted as nodes of the GNN, and the connection relationships between them are abstracted as edges between nodes. The edge weights are initialized according to the correlation degree to construct the distribution network topology graph. Then, the monitoring data in the standardized dataset are mapped to the corresponding nodes according to the unique ID of the equipment / link. Each node is assigned a multi-dimensional feature vector to form a spatial feature graph with data features. Finally, the feature graph is input into the GNN model. The feature information of each node and its neighboring nodes is aggregated through graph convolutional layers to learn the spatial feature correlation between nodes. The aggregation results are then analyzed by fully connected layers to identify and quantify spatial dependencies such as fault propagation of adjacent lines, regional equipment load coordination, and signal transmission between links and terminals. Finally, these patterns are transformed into quantifiable spatial correlation features.
[0094] In the above embodiments, by extracting the temporal dependencies of the distribution network in the time dimension based on a standardized dataset to generate temporal correlation features, and by combining the network topology to extract spatial dependencies to generate spatial correlation features, a deep analysis of the spatiotemporal coupling of the distribution network operation status is achieved, making the final result more consistent with the actual situation.
[0095] In an exemplary embodiment, the defect type of the distribution network is determined based on the preliminary similarity value, temporal correlation features, and spatial correlation features, including: calculating the temporal correlation feature similarity value between the temporal correlation features and the sample temporal correlation features of each defect type in the preset defect library, and the spatial correlation feature similarity value between the spatial correlation features and the sample spatial correlation features of each defect type in the preset defect library; assigning preset weights to the preliminary similarity value, the temporal correlation feature similarity value, and the spatial correlation feature similarity value, and performing weighted summation to obtain the final similarity of each defect type of the distribution network; and identifying the defect type with a final similarity greater than a preset defect threshold as the defect type of the distribution network.
[0096] In practice, the temporal correlation feature similarity value between the temporal correlation feature and the temporal correlation feature of each defect type in the preset defect library, and the spatial correlation feature similarity value between the spatial correlation feature and the spatial correlation feature of each defect type in the preset defect library are calculated respectively.
[0097] The similarity values of temporal correlation features and spatial correlation features can be calculated using cosine similarity or other similarity calculation methods. This application does not limit the specific methods used in this embodiment.
[0098] Preset weights are assigned to the initial similarity value, the temporal correlation feature similarity value, and the spatial correlation feature similarity value, and weighted summation is performed to obtain the final similarity of each defect type in the distribution network. The calculation formula is shown in formula (3).
[0099] (3)
[0100] Where S is the final similarity, This is the preliminary similarity value. This represents the similarity value of time-series related features. This represents the spatial correlation feature similarity value. , , These are the weighting coefficients, and The final similarity value is calculated using values ranging from 0.2 to 0.6.
[0101] Defect types with a final similarity greater than a preset defect threshold are considered as defect types in the distribution network.
[0102] In some embodiments, if the final similarity values of multiple defect types are all greater than a preset defect threshold, it is determined that there are all defects in the distribution network that exceed the preset defect threshold.
[0103] In the above embodiments, an intelligent judgment mechanism based on a preset defect threshold is used to automatically identify and output defect types whose final similarity exceeds the threshold. This fusion identification method effectively overcomes the limitations of single-feature or single-dimensional analysis, significantly improves the accuracy, robustness, and generalization ability of defect identification, and can accurately distinguish defect types that are similar in similarity but fundamentally different, greatly reducing the false positive rate and false negative rate, and providing highly reliable intelligent decision support for the self-healing operation and maintenance of distribution networks.
[0104] In one exemplary embodiment, the defect types include line grounding faults, line short-circuit faults, switch failure to operate defects, switch malfunction defects, communication link interruption defects, equipment overheating defects, transformer fault defects, and load overload defects.
[0105] In practice, defect types include line grounding faults, line short-circuit faults, switch failure to operate, switch malfunction, communication link interruption, equipment overheating, transformer faults, and load overload.
[0106] In some embodiments, the feature index system corresponding to each defect includes temporal features and spatial features. Temporal features include data change trends, abrupt change amplitude, periodic fluctuation period, and duration. Spatial features include device topology location correlation, link connection relationship, regional load distribution correlation, and adjacent device status correlation.
[0107] In the above embodiments, by constructing an expert database for identifying eight common defects, including line grounding faults, line short-circuit faults, switch failure to operate defects, switch malfunction defects, communication link interruption defects, equipment overheating defects, transformer fault defects, and load overload defects, comprehensive coverage and accurate identification of core fault types of primary equipment, secondary equipment, and communication links in the distribution network are achieved.
[0108] In an exemplary embodiment, the overall information data includes distribution master station data, distribution terminal data, and communication link data; the distribution master station data includes the operating parameters of the distribution network, equipment control commands, dispatch commands, and preliminary fault assessment records; the operating parameters include voltage, current, power, and frequency; the distribution terminal data includes switchgear status, instrument transformer monitoring data, equipment temperature data, partial discharge signals, and terminal alarm information; the switchgear status includes opening and closing status and fault tripping records; the communication link data includes communication latency, data packet loss rate, link connectivity status, signal strength, and transmission bandwidth occupancy rate.
[0109] In practice, the overall information data includes distribution master station data, distribution terminal data, and communication link data. Distribution master station data includes the operating parameters of the distribution network, equipment control commands, dispatch commands, and preliminary fault assessment records. Operating parameters include voltage, current, power, and frequency. Distribution terminal data includes switchgear status, instrument transformer monitoring data, equipment temperature data, partial discharge signals, and terminal alarm information. Switchgear status includes opening and closing status and fault trip records. Communication link data includes communication latency, data packet loss rate, link connectivity status, signal strength, and transmission bandwidth occupancy.
[0110] In some embodiments, the data acquisition method combines real-time acquisition with periodic acquisition. Key data such as fault alarm information and switch status change signals are acquired in real time, while data such as operating parameters and normal communication link status are acquired periodically, ensuring that the data comprehensively covers all core operating links of the power distribution system, including power generation, transmission, distribution, and consumption.
[0111] In the above embodiments, this data system breaks through the information silos of traditional single data sources, providing multi-dimensional feature inputs covering electrical quantities, state quantities, environmental quantities and communication quality for subsequent defect identification. It effectively supports comprehensive monitoring from the status of the equipment itself to the quality of communication transmission, and lays a solid data foundation for the mining of temporal evolution patterns and spatial topological correlation analysis based on trajectory analysis, thereby significantly improving the ability to discover hidden defects and the accuracy of identification.
[0112] In an exemplary embodiment, the process of constructing a preset defect library includes: collecting historical global information data when each defect type occurs, and performing data cleaning and standardization to obtain a historical standard dataset; extracting historical time-series feature vectors and historical spatial feature vectors from the historical standard dataset, and fusing the historical time-series feature vectors and historical spatial feature vectors to obtain the corresponding initial sample feature vectors for each defect type; performing statistical analysis on the multiple initial sample feature vectors corresponding to each defect type according to feature dimensions, and calculating the deviation range of each feature dimension; removing abnormal initial sample feature vectors that exceed the deviation range; performing statistical analysis on the remaining initial sample feature vectors corresponding to each defect type according to feature dimensions to obtain the feature mean of each feature dimension under each defect type, and constructing the corresponding sample feature vector for each defect type; and constructing a preset defect library based on the corresponding sample feature vectors for each defect type.
[0113] In practice, historical full-domain information data for each defect type is collected, and data cleaning and standardization are performed to obtain a historical standard dataset. Historical temporal feature vectors and historical spatial feature vectors are extracted from the historical standard dataset, and these vectors are fused to obtain the initial sample feature vectors for each defect type. Statistical analysis is performed on multiple initial sample feature vectors for each defect type according to their feature dimensions to calculate the deviation range for each feature dimension. Abnormal initial sample feature vectors exceeding the deviation range are removed. Based on the remaining initial sample feature vectors for each defect type, statistical analysis is performed according to their feature dimensions to obtain the feature mean for each feature dimension under each defect type, and this constitutes the sample feature vector for each defect type. Based on the sample feature vectors for each defect type, a pre-defined defect database is constructed.
[0114] In the above embodiments, a standardized knowledge benchmark and intelligent comparison template are established for the identification of distribution network defects. The defect database systematically transforms the operation and maintenance experience of domain experts and historical fault data into a computable and matchable feature index system, enabling the system to perform similarity calculations based on standardized feature vectors, and achieve rapid preliminary matching and type prediction of various defects.
[0115] To illustrate the defect identification method for the power distribution network in this application in detail, an embodiment is described below, and the specific flowchart is as follows: Figure 3 As shown, this application illustrates a defect identification method for a power distribution network in a specific scenario, by way of example.
[0116] First, a comprehensive data collection system is constructed to collect information data from the entire power distribution network. This data is then cleaned and standardized to obtain a standardized dataset. Temporal and spatial feature vectors for each monitored object are extracted from this standardized dataset. These vectors are then fused to obtain the target feature vector.
[0117] Calculate the cosine similarity between the target feature vector and the sample feature vector of each defect type in the preset defect library. The calculation formula is shown in formula (1), and the preliminary similarity value between the target feature vector and each defect type is obtained.
[0118] Based on standardized datasets and the network topology of the distribution network, a trajectory analysis algorithm is used to mine the temporal and spatial correlation features of the standardized data. These features are then incorporated into the preliminary similarity value to calculate the final similarity value. When the final similarity value exceeds a preset threshold, the corresponding defect type is determined.
[0119] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps. It is understood that the steps in different embodiments can be freely combined as needed, and all non-contradictory solutions formed by such combinations are within the scope of protection of this application.
[0120] Based on the same inventive concept, this application also provides a distribution network defect identification device for implementing the above-described distribution network defect identification method. The solution provided by this device is similar to the implementation described in the above method; therefore, the specific limitations in one or more distribution network defect identification device embodiments provided below can be found in the limitations of the distribution network defect identification method described above, and will not be repeated here.
[0121] In one exemplary embodiment, such as Figure 4 As shown, a defect identification device for a power distribution network is provided, comprising: a processing module 401, a fusion module 402, a comparison module 403, an extraction module 404, and a determination module 405, wherein:
[0122] The processing module is used to collect full-domain information data of the power distribution network, perform data cleaning and standardization processing on the full-domain information data, and obtain a standardized dataset.
[0123] The fusion module is used to extract the temporal feature vector and spatial feature vector from the standardized dataset, and fuse the temporal feature vector and the spatial feature vector to obtain the target feature vector.
[0124] The comparison module is used to compare the target feature vector with the sample feature vector of each defect type in the preset defect library to obtain a preliminary similarity value.
[0125] The extraction module is used to extract the temporal correlation features and spatial correlation features of the power distribution network based on the standardized dataset and the network topology of the power distribution network.
[0126] The determination module is used to determine the defect type of the distribution network based on the preliminary similarity value, the temporal correlation feature, and the spatial correlation feature.
[0127] Each module in the aforementioned power distribution network defect identification device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.
[0128] In one exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 5 As shown, the computer device includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The input / output interface is used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, Near Field Communication (NFC), or other technologies. When the computer program is executed by the processor, it implements a defect identification method for a power distribution network.
[0129] The display unit of this computer device is used to form a visually visible image and can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of this computer device can be a touch layer covering the display screen, or buttons, a trackball, or a touchpad set on the casing of the computer device, or an external keyboard, touchpad, or mouse, etc.
[0130] Those skilled in the art will understand that Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.
[0131] In one exemplary embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:
[0132] Collect full-domain information data of the power distribution network, and perform data cleaning and standardization on the full-domain information data to obtain a standardized dataset;
[0133] Extract the temporal feature vector and spatial feature vector from the standardized dataset, and fuse the temporal feature vector and the spatial feature vector to obtain the target feature vector;
[0134] The target feature vector is compared with the sample feature vector of each defect type in the preset defect library to obtain a preliminary similarity value.
[0135] Based on the standardized dataset and the network topology of the distribution network, the temporal correlation features and spatial correlation features of the distribution network are extracted.
[0136] Based on the preliminary similarity value, the temporal correlation feature, and the spatial correlation feature, the defect type of the distribution network is determined.
[0137] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:
[0138] Collect full-domain information data of the power distribution network, and perform data cleaning and standardization on the full-domain information data to obtain a standardized dataset;
[0139] Extract the temporal feature vector and spatial feature vector from the standardized dataset, and fuse the temporal feature vector and the spatial feature vector to obtain the target feature vector;
[0140] The target feature vector is compared with the sample feature vector of each defect type in the preset defect library to obtain a preliminary similarity value.
[0141] Based on the standardized dataset and the network topology of the distribution network, the temporal correlation features and spatial correlation features of the distribution network are extracted.
[0142] Based on the preliminary similarity value, the temporal correlation feature, and the spatial correlation feature, the defect type of the distribution network is determined.
[0143] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, performs the following steps:
[0144] Collect full-domain information data of the power distribution network, and perform data cleaning and standardization on the full-domain information data to obtain a standardized dataset;
[0145] Extract the temporal feature vector and spatial feature vector from the standardized dataset, and fuse the temporal feature vector and the spatial feature vector to obtain the target feature vector;
[0146] The target feature vector is compared with the sample feature vector of each defect type in the preset defect library to obtain a preliminary similarity value.
[0147] Based on the standardized dataset and the network topology of the distribution network, the temporal correlation features and spatial correlation features of the distribution network are extracted.
[0148] Based on the preliminary similarity value, the temporal correlation feature, and the spatial correlation feature, the defect type of the distribution network is determined.
[0149] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.
[0150] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.
[0151] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.
[0152] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.
Claims
1. A method for defect identification in a power distribution network, characterized in that, The method includes: Collect full-domain information data of the power distribution network, and perform data cleaning and standardization on the full-domain information data to obtain a standardized dataset; Extract the temporal feature vector and spatial feature vector from the standardized dataset, and fuse the temporal feature vector and the spatial feature vector to obtain the target feature vector; The target feature vector is compared with the sample feature vector of each defect type in the preset defect library to obtain a preliminary similarity value. Based on the standardized dataset and the network topology of the distribution network, the temporal correlation features and spatial correlation features of the distribution network are extracted. Based on the preliminary similarity value, the temporal correlation feature, and the spatial correlation feature, the defect type of the distribution network is determined.
2. The method according to claim 1, characterized in that, The process of cleaning and standardizing the global information data to obtain a standardized dataset includes: Remove outliers from the global information data, and use the device ID, timestamp, and data type in the power distribution network as a joint key to delete duplicate data, thus obtaining the deleted global information data. The missing data in the deleted global information data is filled in to obtain the filled global information data. The data format of the filled global information data is normalized, and a standardized dataset of the power distribution network is constructed based on the association mapping relationship between the device ID, timestamp and data type.
3. The method according to claim 1, characterized in that, The extraction of temporal and spatial correlation features of the distribution network based on the standardized dataset and the network topology of the distribution network includes: Extract the time dependencies of the standardized data in the standardized dataset in the time dimension, and generate the time-series correlation features of the distribution network based on the time dependencies; Based on the standardized dataset and the network topology, the spatial dependencies of the power distribution network are extracted, and the spatial correlation features of the power distribution network are generated based on the spatial dependencies.
4. The method according to claim 1, characterized in that, The process of determining the defect type of the distribution network based on the preliminary similarity value, the temporal correlation characteristics, and the spatial correlation characteristics includes: Calculate the temporal correlation similarity value between the temporal correlation feature and the temporal correlation feature of each defect type in the preset defect library, and the spatial correlation similarity value between the spatial correlation feature and the spatial correlation feature of each defect type in the preset defect library; Preset weights are assigned to the preliminary similarity value, the temporal correlation feature similarity value, and the spatial correlation feature similarity value, and a weighted sum is performed to obtain the final similarity of each defect type in the distribution network; The defect types with a final similarity greater than a preset defect threshold are taken as the defect types of the power distribution network.
5. The method according to claim 1, characterized in that, The defect types include line grounding faults, line short-circuit faults, switch failure to operate, switch malfunction, communication link interruption, equipment overheating, transformer faults, and load overload.
6. The method according to claim 1, characterized in that, The global information data includes power distribution master station data, power distribution terminal data, and communication link data; The data from the distribution master station includes the operating parameters of the distribution network, equipment control commands, dispatch commands, and preliminary fault assessment records; the operating parameters include voltage, current, power, and frequency. The power distribution terminal data includes switchgear status, transformer monitoring data, equipment temperature data, partial discharge signals, and terminal alarm information; the switchgear status includes opening and closing status and fault trip records. The communication link data includes communication latency, data packet loss rate, link connectivity status, signal strength, and transmission bandwidth utilization.
7. The method according to claim 1, characterized in that, The construction process of the preset defect library includes: Historical global information data for each defect type was collected, and the data was cleaned and standardized to obtain a historical standard dataset. Extract the historical time-series feature vector and historical spatial feature vector from the historical standardized dataset, and fuse the historical time-series feature vector and the historical spatial feature vector to obtain the corresponding initial sample feature vector for each defect type; For each defect type, the feature vectors of multiple initial samples are statistically analyzed according to feature dimensions, and the deviation range of each feature dimension is calculated. Remove abnormal initial sample feature vectors that exceed the deviation range; perform statistical analysis on the remaining initial sample feature vectors for each defect type according to the feature dimension to obtain the feature mean of each feature dimension under each defect type, and construct the corresponding sample feature vector for each defect type. A pre-defined defect library is constructed based on the sample feature vector corresponding to each defect type.
8. A defect identification device for a power distribution network, characterized in that, The device includes: The processing module is used to collect full-domain information data of the power distribution network, perform data cleaning and standardization processing on the full-domain information data, and obtain a standardized dataset; The fusion module is used to extract the temporal feature vector and spatial feature vector from the standardized dataset, and fuse the temporal feature vector and the spatial feature vector to obtain the target feature vector; The comparison module is used to compare the target feature vector with the sample feature vector of each defect type in the preset defect library to obtain a preliminary similarity value. The extraction module is used to extract the temporal correlation features and spatial correlation features of the power distribution network based on the standardized dataset and the network topology of the power distribution network. The determination module is used to determine the defect type of the distribution network based on the preliminary similarity value, the temporal correlation feature, and the spatial correlation feature.
9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.