Data management method and system for power spatial data

Abstract

The invention discloses a data management method and system for power spatial data. According to the method, on the basis of time-space characteristics and power business logic of power spatial data, a deep learning model is adopted to carry out joint modeling on spatial dimensions, time dimensions and business logic dimensions, and a semantic model constrained by the time-space characteristics and the business logic is generated; extracting multi-level semantic features of the power spatial data through embedding and attention mechanisms by using the semantic model; and generating a global power spatial data knowledge base based on the extracted semantic features. In combination with a global knowledge base and a data cleaning algorithm, abnormal, redundant and conflict conditions in the data are detected, and a preliminary repair suggestion is output; and finally, completing accurate restoration of the data according to a verification result, and generating a governed power space data set. According to the method, the intelligent degree of power space data management is remarkably improved, and technical support is provided for power grid operation optimization and accurate management.

Description

Technical Field

[0001] The present invention relates to the technical field of power space data processing, and specifically to a data management method and system for power space data. Background Art

[0002] Power spatial data is an important basic data for power grid operation and management, covering multi-dimensional information such as equipment status, geographic location, time series, and business logic. In the existing technology, the processing methods for power spatial data mainly focus on traditional data cleaning and rule matching methods, and realize data governance through statistical analysis or some algorithms based on artificial intelligence. These technologies have played a certain role in the efficiency and accuracy of data governance and have been widely used in power system management.

[0003] However, with the increasing complexity of power systems and the rapid growth of data volumes, existing technologies face many limitations in managing multi-source heterogeneous data and deep semantic analysis. Traditional methods often lack a deep understanding of data semantics and are unable to effectively extract logical associations and contextual relationships between data. In addition, the detection and repair of abnormal data, redundant data, and conflicting data rely on static rules, which are not flexible enough, and the processing results are difficult to meet the needs of efficient and accurate governance.

[0004] In response to the above problems, the present invention proposes an innovative method to improve the management effect of power spatial data. Summary of the invention

[0005] The present application provides a data management method and system for electric power spatial data to improve the management effect of electric power spatial data.

[0006] The present application provides a data governance method for power spatial data, including: Based on the spatiotemporal characteristics of power spatial data and power business logic, a deep learning model is used to jointly model the spatial dimension, temporal dimension and business logic dimension of power spatial data to generate a semantic model that includes spatiotemporal characteristics and business logic constraints. Using the constructed semantic model, the power spatial data is feature extracted through embedding and attention mechanisms to generate multi-level semantic features, wherein the multi-level semantic features include the spatial correlation, time series relationship, and business logic constraints of the power spatial data; Using the extracted multi-level semantic features, a semantic graph is constructed through graph database technology; based on the semantic graph, a global power spatial data knowledge base is generated, and the logical associations and contextual relationships between power spatial data are dynamically annotated in the knowledge base; Based on the global power spatial data knowledge base and in combination with a data cleaning algorithm, abnormal data, redundant data and conflicting data in the power spatial data are detected, and preliminary repair suggestions are generated; According to the preliminary repair suggestions, and based on the logical rules generated by the semantic model, a semantic consistency check is performed on the repair contents of the abnormal data, redundant data, and conflicting data; According to the verification result of the semantic consistency check, abnormal data, redundant data and conflicting data are repaired to generate a managed power spatial data set.

[0007] Furthermore, based on the spatiotemporal characteristics of the power spatial data and the power business logic, a deep learning model is used to jointly model the spatial dimension, time dimension and business logic dimension of the power spatial data to generate a semantic model of the spatiotemporal characteristics and business logic constraints, including: Acquire multi-source heterogeneous input data of power spatial data, wherein the multi-source heterogeneous input data includes geographical location, operating status, historical time series data and corresponding business logic rules of power equipment; perform hierarchical preprocessing on the multi-source heterogeneous input data to form a spatial feature vector, a time series feature vector and a logic rule feature vector; The spatial feature vector is modeled based on a graph neural network to generate spatial semantic features that include the physical connection relationship and spatial correlation between devices; the time series feature vector is modeled based on a long short-term memory network to generate time semantic features that reflect the historical operation status of the device and the time series relationship; the logical rule feature vector is modeled based on an attention mechanism combined with rule reasoning to generate business semantic features that describe device operation constraints and event logic dependencies; The spatial semantic features, temporal semantic features and business semantic features are fused through a multi-layer fully connected network to generate a semantic model including spatiotemporal characteristics and business logic constraints.

[0008] Furthermore, the layered preprocessing of the multi-source heterogeneous input data includes: The weighted distance is calculated according to the following formula (1): ; in, Indicates the device and equipment The weighted distance between For equipment and equipment The geographical distance between For equipment operating load; For equipment operating load; For equipment and equipment The power transmission capacity between Equipment and equipment The health status of , where 1 indicates perfect health; is the weight coefficient; Smoothing parameters to avoid division by zero; Based on dynamic weighted distance , using weighted The mean clustering algorithm is used to group the power equipment, wherein the weighted The optimization objective function of the mean clustering algorithm adopts the following formula 2: ; in, is the target cluster number; For the clusters; Indicates The number of devices in a cluster; is the total number of electrical equipment; is the clustering balance penalty coefficient; According to the following formula (3), The device generates a spatial feature vector : ; in, For the The cluster where the device is located The geometric center coordinates of is calculated according to formula 1 Devices and their clusters The weighted distance between the centers of Indicates The number of devices in a cluster; For the Devices and their clusters The weighted distance of all devices within.

[0009] Furthermore, the spatial feature vector is modeled based on the graph neural network to generate spatial semantic features including the physical connection relationship and spatial correlation between devices, including: A dynamic graph representing the topological structure between power devices is obtained, wherein the nodes of the dynamic graph represent power devices, the edges represent the physical connection relationship between the devices, and the weight of the edge is dynamically calculated based on the physical distance, transmission capacity and connection status between the devices. The weight of the edge is calculated using the following formula (4): ; in, Indicates the device and equipment At the moment of time The weight of the edge; For Node and The physical distance between Indicates time Time equipment and equipment The connection status range is , where 1 means full connection and 0 means no connection; For Node and The transmission capacity between , and is the weight coefficient; is the smoothing coefficient; According to the following formula (5), the weight of the edge is normalized; ; in, Indicates the device and equipment At the moment of time The normalized edge weights of ; Representation and Node The set of all connected nodes; Based on the normalized edge weights, a dynamic graph neural network is used to model spatial features, where the update formula for the node representation of each layer of the graph neural network is expressed by the following formula (6): ; in, Representation Node In the Feature representation of the layer; For the The trainable weight matrix of the layer; Representation Node In the Feature representation of the layer; is the activation function; Representation and Node The set of all connected nodes; The spatial features are updated in multiple layers through dynamic graph neural networks to generate spatial semantic features that reflect the spatiotemporal correlation between power equipment.

[0010] Furthermore, the attention mechanism combined with rule reasoning is used to model the logical rule feature vector to generate business semantic features that describe device operation constraints and event logic dependencies, including: Extracting a set of logical rules related to the operation and dispatch of electric power equipment, and generating a feature representation for each rule in the rule set, wherein the feature representation includes the constraint conditions, the scope of application, and the dependency relationship with other rules of the rule; Analyzing the logical rule set using a rule-based reasoning engine to construct a dependency graph between the rules, wherein the nodes of the dependency graph represent the rules and the edges represent the logical dependency strength between the rules; The rule nodes in the rule dependency graph are modeled in combination with the self-attention mechanism, and the priority weight of each rule is calculated according to the dependency strength and constraint scope of the rule; According to the priority weight, the logic rule feature vector is weighted and aggregated to generate business semantic features that describe device operation constraints and event logic dependencies.

[0011] The beneficial effects of the technical solution provided by this application include: (1) Through joint modeling and semantic model construction based on deep learning models, it is possible to deeply analyze the spatiotemporal characteristics and business logic constraints of power spatial data, extract multi-level semantic features, and significantly improve the processing capabilities of complex, multi-source heterogeneous data. (2) Using semantic models and semantic consistency verification mechanisms, abnormal data, redundant data, and conflicting data can be efficiently detected and accurately repaired to ensure the logical consistency and semantic integrity of the repaired content. (3) Through the construction of semantic graphs and the generation of a global knowledge base, dynamic annotation of logical associations and contextual relationships between power spatial data can be achieved, comprehensively improving the relevance and availability of power data governance. (4) Through the construction of semantic graphs and a global knowledge base, semantic support is provided for subsequent power grid operation optimization, precise management, and user queries, greatly improving the efficiency and intelligence of data query and analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 It is a flow chart of a data governance method for power space data provided in the first embodiment of the present application.

[0013] Figure 2 It is a schematic diagram of a data management system for power space data provided in the second embodiment of the present application. DETAILED DESCRIPTION

[0014] Many specific details are described in the following description to facilitate a full understanding of the present application. However, the present application can be implemented in many other ways than those described herein, and those skilled in the art can make similar generalizations without violating the connotation of the present application, so the present application is not limited by the specific implementation disclosed below.

[0015] The first embodiment of the present application provides a data management method for power space data. Figure 1 , which is a schematic diagram of the first embodiment of the present application. Figure 1 A data governance method for power space data is described in detail in the first embodiment of the present application.

[0016] Step S101: Based on the spatiotemporal characteristics of power spatial data and power business logic, a deep learning model is used to jointly model the spatial dimension, time dimension and business logic dimension of power spatial data to generate a semantic model of spatiotemporal characteristics and business logic constraints.

[0017] Step S101 involves joint modeling based on the spatiotemporal characteristics of power spatial data and power business logic through a deep learning model to generate a semantic model that can describe data characteristics and business constraints. Specifically, this step includes the following implementation methods: First, the original power spatial data is preprocessed to meet the input requirements of the deep learning model. Preprocessing includes but is not limited to data cleaning, format standardization, and feature encoding. Cleaning operations are used to remove obvious noise data or incomplete data; format standardization can transform spatial locations, time series, and related power business events into a unified input structure; feature encoding can transform unstructured data (such as power equipment types or business rules) into vector form that can be input into the model through embedding methods.

[0018] After data preprocessing is completed, the various dimensions of power spatial data are parsed. The parsing of spatial dimension data includes converting geographic location information into a coordinate system or topological structure to represent the spatial relationship between power equipment. The parsing of time dimension data includes sampling and segmenting time series data, such as organizing data on power load or equipment operating status by time window. Business logic dimension data needs to be combined with specific power business rules and equipment operation logic to extract key events or operation sequences.

[0019] Next, a deep learning model is designed and trained. The model consists of three parts: a feature extraction module for the spatial dimension, a feature extraction module for the temporal dimension, and a feature extraction module for the business logic dimension. In the spatial dimension, a graph neural network can be used to capture the spatial correlation between power equipment; in the temporal dimension, a long short-term memory network (LSTM) or a Transformer structure can be used to capture the time series pattern of data; in the business logic dimension, a model based on an attention mechanism can be used to model the logical constraints between events. The output features of the above modules are comprehensively modeled through a fully connected network or other fusion mechanisms to form a joint representation.

[0020] The model is trained based on a labeled training data set using supervised learning, semi-supervised learning, or unsupervised learning. During the training process, a variety of loss functions can be used, such as mean square error (MSE) loss for time prediction, categorical cross entropy loss for logical constraints, etc. After training, the model can generate a semantic model that describes the spatiotemporal characteristics of power space data and business logic constraints.

[0021] The generation of semantic models mainly includes parameter fixation and function encapsulation of the model. The generated semantic model can be directly used for subsequent feature extraction and other tasks, and the model structure and parameters can be stored in a callable file format (such as HDF5 or ONNX). As an output, the semantic model can be used to describe and quantify the multidimensional characteristics of power spatial data and provide a basis for subsequent steps. Through this step, the complex characteristics of power spatial data are systematized and modeled, which is convenient for subsequent processing and application.

[0022] Furthermore, based on the spatiotemporal characteristics of the power spatial data and the power business logic, a deep learning model is used to jointly model the spatial dimension, time dimension and business logic dimension of the power spatial data to generate a semantic model of the spatiotemporal characteristics and business logic constraints, including: Acquire multi-source heterogeneous input data of power spatial data, wherein the multi-source heterogeneous input data includes geographical location, operating status, historical time series data and corresponding business logic rules of power equipment; perform hierarchical preprocessing on the multi-source heterogeneous input data to form a spatial feature vector, a time series feature vector and a logic rule feature vector; The spatial feature vector is modeled based on a graph neural network to generate spatial semantic features that include the physical connection relationship and spatial correlation between devices; the time series feature vector is modeled based on a long short-term memory network to generate time semantic features that reflect the historical operation status of the device and the time series relationship; the logical rule feature vector is modeled based on an attention mechanism combined with rule reasoning to generate business semantic features that describe device operation constraints and event logic dependencies; The spatial semantic features, temporal semantic features and business semantic features are fused through a multi-layer fully connected network to generate a semantic model including spatiotemporal characteristics and business logic constraints.

[0023] First, it is necessary to obtain multi-source heterogeneous input data. This data may include the geographic location information of power equipment, the real-time operating status of the equipment, historical time series records, and business logic rules related to power grid operation. Geographic location information can be obtained through the geographic coordinates or topological connection data of the equipment. The operating status may include real-time parameters such as voltage, current, and load. The time series data records the changes in the operating status of the equipment over time, and the business logic rules reflect the operating specifications and dispatching strategies of the power system.

[0024] After acquiring the data, multi-source heterogeneous data needs to be hierarchically preprocessed to form a feature representation suitable for model processing. Spatial dimension data is converted into spatial feature vectors after encoding and normalization. For example, geographic location information can be mapped to coordinate points or topological matrices. Time dimension data is segmented and serialized through time windows to generate time series feature vectors. These vectors capture the regularity of device operating status changes over time. Business logic data is converted into logical rule feature vectors through rule parsing and embedding technology, which retains the causal relationship and operational constraints between events.

[0025] In the modeling stage, the spatial feature vectors are processed by graph neural networks. Graph neural networks are modeled based on the physical connection relationship and topological structure of the equipment, and spatial semantic features are extracted from them. These features reflect the spatial correlation between power grid equipment, such as direct connection or adjacency between devices. Time series feature vectors are modeled by long short-term memory networks (LSTMs). LSTMs are good at processing data with strong time dependence and can capture the historical laws and future trends of equipment operating status. For business logic feature vectors, the attention mechanism is combined with rule reasoning. The attention mechanism dynamically assigns the importance weights of the rules, while rule reasoning ensures the derivation of logical features and constraint consistency, thereby generating business semantic features.

[0026] After modeling is completed, spatial semantic features, temporal semantic features, and business semantic features are fused through a multi-layer fully connected network. In the fully connected network, each feature is comprehensively represented through weighted processing, and the fusion process highlights the interactive relationship and global significance between the features. The fused features are further nonlinearly activated and dimensionally compressed to finally generate a semantic model that includes both the spatiotemporal characteristics of power data and business logic constraints.

[0027] This method fully captures the complex characteristics of power spatial data through joint modeling of deep learning models, providing a reliable semantic foundation for subsequent data governance, anomaly detection and decision support.

[0028] Furthermore, the layered preprocessing of the multi-source heterogeneous input data includes: The weighted distance is calculated according to the following formula (1): ; in, Indicates the device and equipment The weighted distance between For equipment and equipment The geographical distance between them can be calculated using the geographical coordinates of the devices.

[0029] For equipment operating load; For equipment operating load; For equipment and equipment The power transmission capacity between Equipment and equipment The health status of , where 1 means completely healthy; the health status of the equipment is determined by monitoring the key performance indicators of the equipment, which usually include the equipment's operating history, maintenance records, number of failures, electrical parameters (such as voltage and current stability), etc. The value of the health status is between 0 and 1, where: 1 means the device is completely healthy: the device is in good operating condition, there are no obvious anomalies in the historical data, and all key performance indicators are within the normal range. For example, the device has never failed or damaged, the electrical parameters remain stable, and regular maintenance is completed on time.

[0030] 0 means the device is completely unhealthy: the device has serious problems or is not functioning, such as persistent failures, abnormal electrical parameters, or damage to key components.

[0031] Values ​​between 0 and 1: Reflect the health of the device. For example, if the device occasionally fluctuates in operation, or has slight aging that does not affect its function, a health status of 0.8 or 0.9 can be given; if the device has experienced several failure repairs but is still operating, it can be evaluated as 0.5 or lower.

[0032] is the weight coefficient, obtained through experiments or set according to expert knowledge; Smoothing parameter to avoid division by zero; Based on dynamic weighted distance , using weighted The mean clustering algorithm is used to group the power equipment, wherein the weighted The optimization objective function of the mean clustering algorithm adopts the following formula 2: ; in, is the target cluster number; For the clusters; Indicates The number of devices in a cluster; is the total number of electrical equipment; is the clustering balance penalty coefficient, obtained through experiments or set based on expert knowledge; According to the following formula (3), The device generates a spatial feature vector : ; in, For the The cluster where the device is located The geometric center coordinates of is calculated according to formula 1 Devices and their clusters The weighted distance between the centers of Indicates The number of devices in a cluster; For the Devices and their clusters The average weighted distance of all devices within.

[0033] First, we obtain the basic attributes of power devices, such as geographic location, operating load, power transmission capacity, and health status, and use these attributes to calculate the weighted distance between devices. The weighted distance not only considers physical attributes (such as geographic distance and transmission capacity), but also integrates operating characteristics (such as load differences and health status), providing a comprehensive device association measure.

[0034] On this basis, the weighted k-means clustering algorithm is used to group the devices. The optimization goal of the algorithm is to balance the closeness of the devices within the cluster and the balance between clusters, ensuring that the grouping results can reflect the close association of the devices while avoiding the situation where a single cluster is too sparse or dense.

[0035] Finally, a spatial feature vector is generated for each device. The spatial feature vector consists of the geometric center coordinates, weighted distance, and average weighted distance, which fully describes the spatial location of the device and its relationship with the cluster center and other devices. The design of this feature vector provides high-quality input data for subsequent modeling and analysis.

[0036] Furthermore, the spatial feature vector is modeled based on the graph neural network to generate spatial semantic features including the physical connection relationship and spatial correlation between devices, including: A dynamic graph representing the topological structure between power devices is obtained, wherein the nodes of the dynamic graph represent power devices, the edges represent the physical connection relationship between the devices, and the weight of the edge is dynamically calculated based on the physical distance, transmission capacity and connection status between the devices. The weight of the edge is calculated using the following formula (4): ; in, Indicates the device and equipment At the moment of time The weight of the edge; this value is calculated by Formula 4 and is a dynamically updated quantity used to measure the real-time association between two devices. A larger weight indicates that the relationship between the two devices is more important.

[0037] For Node and The physical distance between Indicates time Time equipment and equipment The connection status range is , where 1 indicates full connection and 0 indicates no connection; connection status is a value between 0 and 1 that indicates the degree of connectivity between devices. It can be obtained by monitoring the operating status of the devices. For example, if two devices are fully connected (such as a circuit breaker is closed), then = 1; if there is no continuity (such as a circuit breaker is open), then = 0. When the connection quality is between the two, a reasonable value can be given according to the operating status of the device, such as the current transmission performance or stability, such as 0.5 for partial connectivity or medium connection quality.

[0038] For Node and The transmission capacity between , and are weight coefficients, which can be determined through experiments or optimization methods. For example, if the physical distance has little effect on device association, you can set For a lower value, it can be increased when it is more dependent on the transmission capacity. .

[0039] This parameter is used to avoid the denominator being zero and is usually set to a very small positive number. Its function is to improve the stability of the calculation.

[0040] According to the following formula (5), the weight of the edge is normalized; ; in, Indicates the device and equipment At the moment of time The normalized edge weights of ; Representation and Node The set of all connected nodes; Formula (5) is used to normalize the edge weights and calculate the normalized edge weights . Normalization is done by placing the device and equipment The edge weight Divide by nodes The normalization operation ensures that the sum of the weights of all adjacent edges is 1, thus distributing a reasonable weight ratio in the graph neural network calculation. For example, if the node Connected to 3 nodes, their edge weights are , after normalization they still retain their relative proportions, but their sum is 1.

[0041] Based on the normalized edge weights, a dynamic graph neural network is used to model spatial features, where the update formula for the node representation of each layer of the graph neural network is expressed by the following formula (6): ; in, Representation Node In the Feature representation of the layer; For the The trainable weight matrix of the layer; Representation Node In the Feature representation of the layer; is the activation function; Representation and Node The set of all connected nodes; Formula (6) is the node update formula of the dynamic graph neural network, which is used to update the node features layer by layer. Layer, Node The feature representation is the weighted sum of all its adjacent node features, by normalizing the edge weight Weighted. Weight matrix It is a trainable parameter used to learn the linear transformation of node features. (such as ReLU or Leaky ReLU) introduces nonlinearity to improve the expressiveness of the network.

[0042] The spatial features are updated in multiple layers through dynamic graph neural networks to generate spatial semantic features that reflect the spatiotemporal correlation between power equipment.

[0043] During the entire modeling process, the dynamic graph neural network gradually captures the spatial correlation and dynamic spatiotemporal characteristics between power equipment through multi-layer node updates. For example, in a transmission network, if transformer A is physically close to circuit breaker B and is currently in a closed state, the edge weight will be higher, and the features of A and B will have a greater mutual influence during the updating process.

[0044] Step S102: Using the constructed semantic model, feature extraction is performed on the power space data through embedding and attention mechanisms to generate multi-level semantic features, wherein the multi-level semantic features include the spatial correlation, time series relationship, and business logic constraints of the power space data.

[0045] The implementation of step S102 involves using the semantic model generated in step S101 to extract features from the power spatial data through embedding and attention mechanisms to generate multi-level semantic features including spatial correlation, time series relationship and business logic constraints. The specific implementation is as follows.

[0046] First, based on the generated semantic model, the input data for feature extraction is initialized. The input of power spatial data includes feature vectors encoded in its spatial dimension, time dimension and business logic dimension, which are further embedded based on the semantic model. Data embedding in the spatial dimension can adopt a spatial embedding algorithm based on positional relationship, such as mapping geographic location into vector form to quantify the spatial association between devices; data embedding in the time dimension maps data into the time feature space through a time series feature processing module, such as time window encoding or timestamp embedding method; data embedding in the business logic dimension uses business rules and event sequences to convert them into structured features that can be input into deep learning models.

[0047] After the initial embedding is completed, feature extraction relies on the attention mechanism to capture the multi-dimensional correlation of the embedded vector. Specifically, the attention module of the semantic model is used to weight the input embedded data to extract important interactive characteristics between the data. The attention mechanism in the spatial dimension can dynamically assign weights between various power devices and highlight pairs of devices with strong correlations. The attention mechanism in the time dimension analyzes the correlation between different time points in the time series data to capture key time-dependent features. The attention mechanism in the business logic dimension dynamically assigns weights according to the logical relationship of events, thereby extracting important patterns in business logic constraints.

[0048] Through the embedding and attention mechanism, the obtained feature vector is input into the feature fusion layer of the deep learning model. The fusion layer combines features of different dimensions into a unified multi-level semantic feature representation through nonlinear activation functions (such as ReLU or Leaky ReLU) and fully connected layers. The multi-level semantic features contain spatial correlation, time series relationship, and business logic constraints, which can fully reflect the inherent characteristics and complex relationships of power space data.

[0049] After feature extraction, the multi-level semantic features are further formatted for subsequent steps. The formatting process includes normalization and dimension compression of feature vectors to optimize storage and computing efficiency. These feature data can be stored in the form of matrices or tensors as the basis for the next step of semantic graph construction. Through this series of operations, step S102 lays a solid semantic foundation for the deep governance of power spatial data and provides high-quality input for subsequent data analysis and processing.

[0050] Furthermore, the constructed semantic model is used to extract features from the power space data through embedding and attention mechanisms to generate multi-level semantic features, including: Embed the spatial dimension of power spatial data to convert the geographical location information, connection relationship and physical topology characteristics of the equipment into high-dimensional feature representation that can be processed by the model; The time dimension of power spatial data is serialized and embedded, and the historical records and time dependencies of equipment operation status are mapped into continuous time feature representation to capture the impact of key time points; The business logic dimension of power spatial data is processed by rule embedding, and the logical dependency relationship between power business rules and events is converted into a logical feature vector, retaining the causal relationship between events; The above embedded features are weighted through the attention mechanism, and the weight of each feature is dynamically assigned. Key features are extracted according to the importance of spatial correlation, time series patterns and business rules, generating multi-level semantic features including spatial correlation, time series relations and business logic constraints.

[0051] First, it is necessary to embed the various dimensions of the power spatial data to generate a high-dimensional feature representation suitable for model calculation. For data in the spatial dimension, the focus of the embedding process is to encode the geographical location information of the equipment, the physical connection relationship between the power equipment, and the topological characteristics of the power grid into high-dimensional features that can be used by the model. Specifically, the geographical location of the equipment can be converted into a coordinate vector, and the connection relationship between the equipment can be represented by an adjacency matrix or a topological graph. Physical topological characteristics, such as the degree and connectivity of the node, can be further converted into embedded features using specific encoding techniques, thereby providing the model with association information between devices.

[0052] The embedding process of the time dimension focuses on the time series information of the equipment's operating status. The historical record data of the equipment, such as load, voltage, or fault status, can be serialized into time series features and segmented using time window technology. These segmented data are then mapped into continuous time features to capture the dependencies and trends between time points. For example, timestamp embedding or position encoding methods can be used to enable the model to identify the impact of key time points on data patterns.

[0053] For data in the business logic dimension, the core of the embedding process is to encode the logical dependencies of power business rules and events into logical feature vectors. Business rules may include dispatch processes, fault handling rules, or maintenance plans, which are converted into logical expressions through semantic parsing technology and further mapped into feature vectors. The logical dependencies of events, such as causal chains or priorities, are integrated into the logical feature vectors through rule reasoning to ensure that the model can identify causal relationships and constraints between events.

[0054] After the above embedding process is completed, the generated features are weighted through the attention mechanism. The role of the attention mechanism is to dynamically adjust the weight of each feature and allocate resources according to its relative importance to the current task or goal. For example, for some tasks, spatial features may be more important than temporal features, while for other tasks, logical rules may play a key role. The attention mechanism assigns appropriate weights to each feature by learning the correlation between features, thereby highlighting the contribution of key features to the model.

[0055] Finally, the multi-level semantic features generated by embedding and attention mechanisms include spatial correlation, time series relationship, and business logic constraints. These features comprehensively represent the multi-dimensional information of power spatial data, providing high-quality input support for subsequent semantic modeling, analysis, and data governance. This method ensures that the complex characteristics of power spatial data in different dimensions can be fully captured and effectively utilized.

[0056] Step S103: Utilize the extracted multi-level semantic features to construct a semantic graph through graph database technology; generate a global power space data knowledge base based on the semantic graph, and dynamically annotate the logical associations and contextual relationships between power space data in the knowledge base.

[0057] Step S103 involves using the multi-level semantic features extracted in step S102 to construct a semantic graph through graph database technology, and generating a global power space data knowledge base based on the semantic graph, while dynamically annotating the logical associations and contextual relationships between power space data in the knowledge base. The specific implementation is as follows: First, the extracted multi-level semantic features need to be formatted to make them suitable for the input requirements of the graph database. This includes hierarchical arrangement of feature vectors, clarifying the node types of the data (such as power equipment, geographic location, time nodes, etc.) and the types of edges between them (such as physical connection relationships between devices, time series associations of events, business logic constraints, etc.). These nodes and edges are labeled with unique identifiers to ensure their uniqueness and identifiability when building a graph database.

[0058] Next, the semantic graph is constructed using graph database technology. Graph database technology can be implemented using existing graph database platforms (such as Neo4j, JanusGraph, etc.), and its construction steps include the definition of nodes and edges, the addition of attributes, and the creation of relationships. In the node definition, specific attributes are assigned to each node type. For example, the power equipment node contains attributes such as equipment ID, equipment type, and geographic location, and the event node contains attributes such as event type, timestamp, and related equipment. The definition of the edge is based on the relationship type between nodes, such as "connected to", "occurred in", or "controlled by", etc. Each edge type can be attached with weight or directional information to reflect the strong or weak association or causal relationship between the data.

[0059] In the process of building a semantic graph, it is necessary to dynamically map multi-level semantic features into specific elements in the graph structure. For example, spatial correlation features are used to define the spatial adjacency relationship between device nodes, time series relationships are used to connect time nodes with related events, and business logic constraints are used to establish logical connections between business rule nodes and device nodes or event nodes. This mapping process can be implemented through scripts or automated tools to ensure that all data elements are correctly reflected in the graph structure.

[0060] After the initial construction of the semantic graph is completed, a global power spatial data knowledge base is generated based on the graph. The generation of the knowledge base is based on the semantic graph, and further integrates external data sources (such as power grid topology data, historical operation data, business rule documents, etc.) to supplement and improve the existing graph information. In the process of knowledge base generation, logical rules and contextual relationship rules can be defined to dynamically annotate the logical associations and contextual information between data. For example, the rule engine can be used to dynamically annotate the current flow relationship based on the electrical characteristics between devices, or dynamically generate a contextual description of the event chain based on the time sequence of events.

[0061] The dynamic annotation mechanism of the knowledge base is implemented through real-time updates and reasoning engines. For example, when new data is input or existing data changes, the knowledge base can automatically update the attributes or relationships of related nodes and edges, thereby ensuring that the logical associations and contextual relationships of the data are always kept up to date. This dynamic annotation mechanism relies on the real-time computing capabilities of the graph database and the logical rule support provided by the semantic model.

[0062] Through the above steps, the construction of semantic graph and the generation of global power spatial data knowledge base provide comprehensive, accurate and efficient support for subsequent data analysis and processing. The method makes full use of the advantages of multi-level semantic features and graph database technology to achieve deep semantic analysis and structured organization of power spatial data.

[0063] Furthermore, the method of using the extracted multi-level semantic features to construct a semantic graph through graph database technology includes: The extracted multi-level semantic features are divided into feature sets of nodes and edges, where node features include the spatial location, time state and business logic attributes of power equipment, and edge features include the physical connection relationship between devices, the time dependency relationship between events, and the logical rule constraint relationship; The graph structure is defined based on node features and edge features, where nodes represent power equipment or key events, and edges represent associations between equipment or logical dependencies between events; Create a storage structure for nodes and edges in the graph database, and attach attributes to each node and edge, including spatial characteristics, temporal characteristics, and contextual information of business rules; According to the spatiotemporal changes of power spatial data, the nodes and edges in the graph structure are dynamically updated to ensure that the semantic graph can reflect changes in equipment status and adjustments to business logic in real time.

[0064] First, the extracted multi-level semantic features need to be systematically divided. Multi-level semantic features include three main dimensions: space, time, and logic. The node feature set mainly covers the core attributes related to power equipment, such as the geographical location, power parameters, real-time status, and business logic constraints of the equipment. This information jointly describes the spatial location, time status, and operational attributes of the equipment. The edge feature set includes the physical connection relationship between devices, the time dependency between events, and logical rule constraints. These edge features are used to describe the correlation and logical interaction between nodes.

[0065] Based on the divided node and edge features, the graph structure for the semantic graph is defined. The nodes of the graph structure can represent power equipment or key events, such as equipment nodes such as transformers, switches, circuit breakers, or event nodes such as system scheduling and abnormal alarms. The edges represent the relationship between nodes, such as the power transmission path in the physical topology, the time sequence dependency in the event trigger chain, and the logical association in the business rule constraints. In this way, the graph structure can fully express the multi-dimensional association of power space data.

[0066] In a graph database, nodes and edges are stored in a defined graph structure. Each node stores its attribute information, such as geographic coordinates, device category, and real-time operating status; each edge stores its associated context information, such as the physical distance between devices, transmission capacity, and time trigger intervals and logical dependency rules between events. These attributes and context information are attached to nodes and edges, so that each element has a complete semantic description that can be traced and queried.

[0067] To ensure that the semantic graph can adapt to the dynamic changes of power space data, the nodes and edges in the graph structure need to be updated in real time. For example, when the state of the equipment changes (such as from running state to fault state), the corresponding node attributes need to be updated immediately; when the physical topology is adjusted or the event chain is added, the corresponding edge attributes or connection relationships need to be modified synchronously. Through these update operations, the semantic graph can remain consistent with the actual power space data and reflect the latest information in a timely manner when the data changes dynamically.

[0068] Ultimately, this semantic graph built on graph database technology can provide an efficient structured representation for power spatial data. The semantic graph not only contains multi-dimensional information about power equipment and events, but also reveals complex logical associations and contextual dependencies through the relationships between nodes and edges, providing a powerful support tool for power data governance, analysis, and decision-making.

[0069] Furthermore, the generation of a global power space data knowledge base based on the semantic graph includes: Extracting semantic information of nodes and edges from the semantic graph, where the semantic information of nodes includes the geographical location, power operation parameters and real-time status of the equipment, and the semantic information of edges includes the physical connection between devices, the time sequence relationship between events and the strength of association; Based on the extracted node and edge information, the core structure of the knowledge base is defined, each node is mapped to an entity in the knowledge base, and the edge is mapped to the association relationship between entities, while adding contextual attributes related to device operation, time logic, and business rules; The data in the semantic graph is stored in the knowledge base, using a scalable storage format to support large-scale data operations, and a dynamic maintenance node and edge update mechanism to ensure that the knowledge base can synchronize changes in the semantic graph in real time.

[0070] First, it is necessary to extract the complete semantic information of nodes and edges from the constructed semantic graph. The semantic information of nodes should cover the geographical location, power operation parameters and real-time status of power equipment. For example, each node can represent a specific power equipment, including its location in the power grid topology, current load, voltage level, and operating health status. The semantic information of edges is used to describe the relationship between nodes, including the physical connection characteristics between devices, such as the transmission capacity and length of power lines, as well as the temporal order relationship and logical association strength between events, such as the triggering dependency of abnormal events.

[0071] Based on the extracted node and edge information, the core structure of the global power spatial data knowledge base is defined. The entities in the knowledge base correspond to nodes in the semantic graph, such as transformers, switches, circuit breakers, or specific scheduling events. The associations between entities are mapped from the edges in the semantic graph, such as the power transmission path between devices or the trigger logic between events. Each entity and association is attached with rich contextual attributes, including spatial characteristics (such as the geographic coordinates of the device), temporal logic (such as the time sequence of events), and business rules (such as scheduling dependencies or fault impact range). These contextual attributes ensure that the knowledge base can comprehensively represent the multi-dimensional correlation of power spatial data.

[0072] When the data in the semantic graph is stored in the knowledge base, a scalable storage format is required to support large-scale power data operations. For example, the knowledge base can be implemented based on a relational database or a graph database to ensure that complex associations can be quickly queried on the basis of storing nodes and edges. In order to adapt to the dynamic changes in actual power spatial data, the information of nodes and edges in the knowledge base needs to have a dynamic maintenance mechanism. For example, when the operating status of a device changes, the attributes of the corresponding entity in the knowledge base should be updated in real time; when the power grid topology is adjusted, the newly added or removed devices and their connection relationships should be synchronously reflected in the knowledge base. This dynamic update mechanism ensures that the content of the knowledge base is always consistent with the semantic graph.

[0073] The resulting global power spatial data knowledge base provides comprehensive semantic data support and can be queried, reasoned, and analyzed through standardized interfaces. The knowledge base not only reflects the current status of power equipment and events, but also reveals the complex logic and associations of power spatial data through its structured semantic representation, providing a reliable foundation for subsequent power data governance, scheduling optimization, and decision support. Through this approach, the global power spatial data knowledge base can manage and utilize power spatial data in an efficient, dynamic, and scalable manner.

[0074] Step S104: Based on the global power space data knowledge base and in combination with a data cleaning algorithm, abnormal data, redundant data and conflicting data in the power space data are detected, and preliminary repair suggestions are generated.

[0075] The implementation of step S104 involves detecting abnormal data, redundant data and conflicting data in the power space data based on the global power space data knowledge base and combining the data cleaning algorithm, and generating preliminary repair suggestions. Specifically, this step first extracts the target data set to be cleaned from the global power space data knowledge base, which usually includes all power data involving spatial dimensions, time dimensions and business logic dimensions. These data will be classified and sorted according to their node types (such as device nodes, event nodes) and edge types (such as association relationships, time series relationships) for subsequent processing.

[0076] For the sorted data, the target data set is processed in combination with the data cleaning algorithm. Abnormal data detection is mainly achieved by setting contextual constraints and logical rules. For example, for power equipment operation data, abnormal values ​​can be identified by monitoring whether the equipment parameters exceed the normal working range; for time series data, missing data or time logic conflict data can be identified by analyzing the continuity of time nodes; for business logic, the business rules defined in the knowledge base can be used to detect operations or states that violate the rules. In specific implementations, abnormal data detection can adopt a rule-based approach, or it can be combined with statistical analysis models or machine learning algorithms to improve the robustness and flexibility of detection.

[0077] Redundant data detection focuses on identifying records that are stored repeatedly or have the same content but different sources. For example, by comparing the topological structure between power equipment nodes, it is possible to detect whether there are duplicate nodes; or by annotating the nodes and edges in the semantic graph, it is possible to confirm whether there are duplicate logical relationships. The detected redundant data will be marked as candidate data for further confirmation during subsequent repairs.

[0078] Conflict data detection aims to identify contradictions between data, such as data from multiple sources having different values ​​under the same context. When detecting conflicting data, reasoning can be performed based on the contextual relationship rules in the knowledge base, such as whether the logical association between devices conforms to the topological characteristics of the power system, or whether the sequence of time series events conforms to the power dispatch rules. In complex conflict situations, the reasoning ability of the semantic model can also be used to analyze the cause of the conflict.

[0079] After the detection is completed, preliminary repair suggestions are generated based on the detection results. Repair suggestions include specific operational instructions, such as recommended correction values ​​for abnormal data, suggestions for merging or deleting redundant data, and preferred processing solutions for conflicting data. When generating repair suggestions, a weight value can be added to each repair suggestion based on the importance of the data or the credibility of the source for reference in subsequent processing links. This process is usually implemented through a rule engine or decision tree model to ensure that the generated repair suggestions are logically consistent and operational.

[0080] Through the above steps, this stage effectively detects and marks quality issues in power spatial data and generates preliminary repair suggestions, laying a solid foundation for semantic consistency verification and final repair in subsequent steps.

[0081] Step S105: According to the preliminary repair suggestion and the logic rules generated by the semantic model, a semantic consistency check is performed on the repair contents of the abnormal data, redundant data and conflicting data.

[0082] The implementation of step S105 involves performing semantic consistency checks on the repair contents of abnormal data, redundant data, and conflicting data based on preliminary repair suggestions and using logical rules generated by the semantic model to ensure that the repair process is consistent with the inherent logic and business semantics of the power space data.

[0083] First, the preliminary repair suggestions generated in step S104 are input into the semantic consistency verification module, the core of which relies on the logical rule base generated by the semantic model. The logical rule base includes a multi-level rule set in the spatial dimension, the temporal dimension, and the business logic dimension, such as the physical connection constraint rules between power equipment, the temporal relationship rules of power dispatching events, and the causal association rules between business operations and states. These rules are generated by the semantic model in the joint modeling and feature extraction stage, and can accurately describe the logical association and contextual relationship of power space data.

[0084] When performing semantic consistency verification, for each repair suggestion, the repair content is first matched with the existing data in the knowledge base to confirm whether the repair content conforms to the existing context. For example, for abnormal data repair suggestions, the verification module will verify whether the repair value falls within the normal working range marked in the knowledge base; for redundant data merging suggestions, it will confirm whether the merged data conforms to the requirements of the device topology or business logic; for the preferred solution for conflicting data, the credibility or consistency of the selected data will be evaluated through logical rule reasoning.

[0085] The verification process adopts a layer-by-layer verification method. First, the rules of the spatial dimension are verified, such as whether the physical connection relationship between power equipment is complete, and whether the repaired spatial data introduces new conflicts or inconsistencies. Then the rules of the time dimension are verified to confirm whether the events in the time series conform to the laws of power dispatch or operation, such as whether the sequence of events after the repair conforms to the time dependency. Finally, the rules of the business logic dimension are verified to ensure that the repair content conforms to the causal relationship and operation specifications of the power business. For example, if the repair value of an event causes inconsistency with the preceding or following operations in the business process, the repair content will be marked as inconsistent.

[0086] After completing the semantic consistency check, the verification module will output the verification results for each repair suggestion, including three states: "passed", "needs adjustment" or "rejected". If the check passes, the repair content can directly enter the next execution stage; if adjustment is required, the system will feedback the rules or data attributes that need to be improved based on the specific reasons for the failure of the check; if the suggestion is rejected, it means that the repair content seriously conflicts with the semantic rules and is not eligible for execution.

[0087] Through this process, step S105 ensures that the repair suggestions are not only effective for the data itself, but also consistent with the global logic and business requirements of the power space data in terms of overall semantics, thus laying the foundation for the high quality and high reliability of the final data governance results.

[0088] Furthermore, according to the preliminary repair suggestion and the logic rules generated by the semantic model, a semantic consistency check is performed on the repair content of the abnormal data, redundant data and conflicting data, including: Extracting the target data and its context information involved in the preliminary repair suggestion, including the source of the repair data, applicable logical rules and contextual association conditions; Based on the logical rule base generated by the semantic model, the repair content is checked one by one to verify whether the repair data meets the physical constraints of the equipment, time series logic and business operation specifications; For repair content involving multiple data sources or relying on contextual conditions, the logical reasoning module verifies whether the repaired data is consistent with the associated data, including the logical integrity of the event chain and the logical continuity of the device status.

[0089] First, the preliminary repair suggestions need to be parsed in detail to extract the target data and its contextual information. This includes clarifying the specific source of the repair content, such as the acquisition device or data source system of the original data record, and extracting the logical rules and contextual conditions related to the repair data. For example, if the repair suggestion involves the correction of an abnormality in the operating status of the equipment, it needs to include the equipment type, the associated physical topology, the related time series records, and the applicable business rules.

[0090] After extracting the target data, the verification process relies on the logical rule base generated by the semantic model, which contains a set of rules for describing the logical constraints of power space data, including but not limited to the physical constraints of equipment, the logical continuity of time series, and the process specifications of business operations. Each rule is associated with a specific contextual condition. For example, the physical connection rules between devices can be defined as part of the topology, while the time logic rules depend on the time sequence of events.

[0091] Next, the repair content is matched with the logical rules in the rule base one by one to verify whether the logical constraints are met. For example, for the repair of equipment operating parameters, it is verified whether it meets the normal operating range of the equipment and whether it is consistent with the status of adjacent equipment; for the repair of missing or abnormal values ​​in the time series, it is verified whether the repaired value is consistent with the trend and dependency of the time series, such as whether it conforms to the law of value changes between time points or the time sequence of event triggering.

[0092] When multiple data sources are involved or complex contextual conditions need to be considered for repair content, the verification process further calls the logical reasoning module. This module compares the repair data with the associated data through semantic reasoning to verify whether it is consistent with the contextual information. For example, in the verification of the event chain, the reasoning module needs to verify whether the repair content maintains logical integrity, such as whether the sequence of events after the repair conforms to the causal relationship; in the verification of the device status, it is necessary to ensure that there is no conflict between the repaired status and the physical topology and operation rules.

[0093] If the repair content is found to be inconsistent with the logical rules or context information during verification, the repair content will be marked as verification failure, and the specific failure reason and conflict point will be recorded, such as violation of specific physical constraints or logical rules. This information will be passed to the repair module as feedback to further adjust the repair suggestions.

[0094] The entire verification process ensures that the repaired data can be logically consistent with the global power space data, complying with physical and business constraints and supporting subsequent data analysis and governance operations.

[0095] Step S106: According to the verification result of the semantic consistency check, abnormal data, redundant data and conflicting data are repaired to generate a managed power space data set.

[0096] The implementation of step S106 involves performing a final repair operation on abnormal data, redundant data, and conflicting data based on the verification result of the semantic consistency check in step S105, thereby generating a governed power space data set. Specifically, this step strictly combines the repair operation with the semantic logic, contextual relationship, and business rules of the power space data to ensure the accuracy and reliability of the repair result.

[0097] First, according to the verification results of the semantic consistency check, the repair content is classified and processed. For the repair suggestions that pass the verification, the repair operation is directly executed to update the repair value or the corrected data to the target data set. For example, for the repair of abnormal data, the original abnormal value is directly replaced with the corrected value that passes the verification; for the repair of redundant data, the most credible record is retained through merging or deduplication operations; for the repair of conflicting data, the preferred solution is used to replace the conflicting item and remove other conflicting values.

[0098] For repair suggestions that need adjustment, the system will improve the repair content based on the adjustment feedback provided by the verification module. This usually includes adjusting the range of repair values, optimizing the strategy for data merging, or re-evaluating the preferred rules for conflicting data. For example, if a time series repair suggestion is marked as requiring adjustment because it does not meet the time continuity rule, the repair value can be recalculated to meet the dependencies of the time logic. The improved repair suggestion needs to pass the consistency check again to ensure that the adjusted repair content meets all semantic rules.

[0099] For the repair content that is rejected, the system will mark it as unrepaired due to its irreconcilable conflict with the semantic rules, and record the conflict for subsequent manual intervention or further analysis. Such data usually needs to be judged by domain experts in combination with actual business scenarios.

[0100] After all repair operations are completed, the system will uniformly format and standardize the repaired data set. Formatting includes field verification, type conversion, and unified encoding of the repaired data to ensure the structural consistency and availability of the data set. At the same time, the system will generate a repair log, which records the original data, repair content, verification results, and final repair value of each repair operation in detail, providing a basis for subsequent auditing and verification.

[0101] Finally, the repaired data set is stored as a governed power spatial data set. This data set not only eliminates abnormal, redundant, and conflicting data, but also ensures the logical and contextual integrity of the data through semantic consistency verification. The governed data set can be directly used for power grid operation optimization, business analysis, and other subsequent applications, thereby significantly improving the quality and utilization value of the data. Through this series of rigorous repair operations, step S106 provides a reliable foundation for the efficient governance and in-depth application of power spatial data.

[0102] In the above-mentioned embodiment, a data management method for electric power space data is provided. Correspondingly, the present application also provides a data management system for electric power space data. Figure 2 , which is a schematic diagram of an embodiment of a data management system for power space data of the present application. Since this embodiment, i.e., the second embodiment, is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. The system embodiment described below is only illustrative.

[0103] A data management system for electric power spatial data provided in the second embodiment of the present application includes: A modeling unit 201 is used to jointly model the spatial dimension, time dimension and business logic dimension of the power spatial data using a deep learning model based on the spatiotemporal characteristics of the power spatial data and the power business logic, and generate a semantic model of the spatiotemporal characteristics and business logic constraints; A generating unit 202 is used to extract features of the power space data by using the constructed semantic model through embedding and attention mechanism to generate multi-level semantic features, wherein the multi-level semantic features include spatial correlation, time series relationship and business logic constraints of the power space data; The construction unit 203 is used to construct a semantic graph by using the extracted multi-level semantic features through the graph database technology; based on the semantic graph, a global power space data knowledge base is generated, and the logical associations and contextual relationships between the power space data are dynamically marked in the knowledge base; A detection unit 204, configured to detect abnormal data, redundant data and conflicting data in the power space data based on the global power space data knowledge base and in combination with a data cleaning algorithm, and to generate preliminary repair suggestions; A verification unit 205 is used to perform semantic consistency verification on the repair contents of abnormal data, redundant data and conflicting data according to the preliminary repair suggestion and the logic rules generated by the semantic model; The repair unit 206 is used to repair abnormal data, redundant data and conflicting data according to the verification result of the semantic consistency check to generate a managed power space data set.

[0104] A third embodiment of the present application provides an electronic device, the electronic device comprising: processor; The memory is used to store a program, and when the program is read and executed by the processor, it executes a data management method for power space data provided in the first embodiment of the present application.

[0105] The fourth embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon. When the program is executed by a processor, a data governance method for power space data provided in the first embodiment of the present application is executed.

[0106] Although the present application is disclosed as above in the form of a preferred embodiment, it is not intended to limit the present application. Any technical personnel in this field may make possible changes and modifications without departing from the spirit and scope of the present application. Therefore, the scope of protection of the present application shall be based on the scope defined by the claims of the present application.

Claims

1. A data management method for power space data, characterized in that: include: Based on the spatiotemporal characteristics of power spatial data and power business logic, a deep learning model is used to jointly model the spatial dimension, temporal dimension and business logic dimension of power spatial data to generate a semantic model that includes spatiotemporal characteristics and business logic constraints. Using the constructed semantic model, the power spatial data is feature extracted through embedding and attention mechanisms to generate multi-level semantic features, wherein the multi-level semantic features include the spatial correlation, time series relationship, and business logic constraints of the power spatial data; Using the extracted multi-level semantic features, a semantic graph is constructed through graph database technology; based on the semantic graph, a global power spatial data knowledge base is generated, and the logical associations and contextual relationships between power spatial data are dynamically annotated in the knowledge base; Based on the global power spatial data knowledge base and in combination with a data cleaning algorithm, abnormal data, redundant data and conflicting data in the power spatial data are detected, and preliminary repair suggestions are generated; According to the preliminary repair suggestions, and based on the logical rules generated by the semantic model, a semantic consistency check is performed on the repair contents of the abnormal data, redundant data, and conflicting data; According to the verification result of the semantic consistency check, abnormal data, redundant data and conflicting data are repaired to generate a managed power spatial data set.

2. The data management method for power space data according to claim 1 is characterized in that: Based on the spatiotemporal characteristics of power spatial data and power business logic, a deep learning model is used to jointly model the spatial dimension, time dimension and business logic dimension of power spatial data to generate a semantic model of spatiotemporal characteristics and business logic constraints, including: Acquire multi-source heterogeneous input data of power spatial data, wherein the multi-source heterogeneous input data includes geographical location, operating status, historical time series data and corresponding business logic rules of power equipment; perform hierarchical preprocessing on the multi-source heterogeneous input data to form a spatial feature vector, a time series feature vector and a logic rule feature vector; The spatial feature vector is modeled based on a graph neural network to generate spatial semantic features that include the physical connection relationship and spatial correlation between devices; the time series feature vector is modeled based on a long short-term memory network to generate time semantic features that reflect the historical operation status of the device and the time series relationship; the logical rule feature vector is modeled based on an attention mechanism combined with rule reasoning to generate business semantic features that describe device operation constraints and event logic dependencies; The spatial semantic features, temporal semantic features and business semantic features are fused through a multi-layer fully connected network to generate a semantic model including spatiotemporal characteristics and business logic constraints.

3. The data management method for power space data according to claim 2 is characterized in that: The step of performing hierarchical preprocessing on the multi-source heterogeneous input data includes: The weighted distance is calculated according to the following formula (1): ； in, Indicates the device and equipment The weighted distance between For equipment and equipment The geographical distance between For equipment operating load; For equipment operating load; For equipment and equipment The power transmission capacity between Equipment and equipment The health status of , where 1 indicates perfect health; is the weight coefficient; Smoothing parameters to avoid division by zero; Based on dynamic weighted distance , using weighted The mean clustering algorithm is used to group the power equipment, wherein the weighted The optimization objective function of the mean clustering algorithm adopts the following formula 2: ； in, is the target cluster number; For the clusters; Indicates The number of devices in a cluster; is the total number of electrical equipment; is the clustering balance penalty coefficient; According to the following formula (3), The device generates a spatial feature vector : ； in, For the The cluster where the device is located The geometric center coordinates of is calculated according to formula 1 Devices and their clusters The weighted distance between the centers of Indicates The number of devices in a cluster; For the Devices and their clusters The average weighted distance of all devices within.

4. The data management method for power space data according to claim 2 is characterized in that: The spatial feature vector is modeled based on a graph neural network to generate a spatial semantic feature including a physical connection relationship and a spatial correlation between devices, including: A dynamic graph representing the topological structure between power devices is obtained, wherein the nodes of the dynamic graph represent power devices, the edges represent the physical connection relationship between the devices, and the weight of the edge is dynamically calculated based on the physical distance, transmission capacity and connection status between the devices. The weight of the edge is calculated using the following formula (4): ； in, Indicates the device and equipment At the moment of time The weight of the edge; For Node and The physical distance between Indicates time Time equipment and equipment The connection status range is , where 1 means full connection and 0 means no connection; For Node and The transmission capacity between , and is the weight coefficient; is the smoothing coefficient; According to the following formula (5), the weight of the edge is normalized; ； in, Indicates the device and equipment At the moment of time The normalized edge weights of ; Representation and Node The set of all connected nodes; Based on the normalized edge weights, a dynamic graph neural network is used to model spatial features, where the update formula for the node representation of each layer of the graph neural network is expressed by the following formula (6): ； in, Representation Node In the Feature representation of the layer; For the The trainable weight matrix of the layer; Representation Node In the Feature representation of the layer; is the activation function; Representation and Node The set of all connected nodes; The spatial features are updated in multiple layers through dynamic graph neural networks to generate spatial semantic features that reflect the spatiotemporal correlation between power equipment.

5. The data management method for power space data according to claim 2 is characterized in that: The attention mechanism combined with rule reasoning is used to model the logical rule feature vector to generate business semantic features that describe device operation constraints and event logic dependencies, including: Extracting a set of logical rules related to the operation and dispatch of electric power equipment, and generating a feature representation for each rule in the rule set, wherein the feature representation includes the constraint conditions, the scope of application, and the dependency relationship with other rules of the rule; Analyzing the logical rule set using a rule-based reasoning engine to construct a dependency graph between the rules, wherein the nodes of the dependency graph represent the rules and the edges represent the logical dependency strength between the rules; The rule nodes in the rule dependency graph are modeled in combination with the self-attention mechanism, and the priority weight of each rule is calculated according to the dependency strength and constraint scope of the rule; According to the priority weight, the logic rule feature vector is weighted and aggregated to generate business semantic features that describe device operation constraints and event logic dependencies.

6. The data management method for power space data according to claim 2 is characterized in that: The spatial semantic features, temporal semantic features and business semantic features are integrated through a multi-layer fully connected network to generate a semantic model including spatiotemporal characteristics and business logic constraints, including: Normalize the spatial semantic features, temporal semantic features, and business semantic features separately to ensure that the range of feature values ​​is consistent and avoid the impact of feature scale differences on model training; Construct a multi-layer fully connected network, where the input of each layer is the fused features of the output of the previous layer, and the network weights adjust the contribution of different feature types through trainable parameters; Regularization constraints are added between features in each fully connected network layer to limit the excessive influence of certain types of features and ensure that the contribution of spatial, temporal, and business features to the final semantic model is balanced. In the last layer of feature fusion, the fused features are nonlinearly mapped through activation functions to generate a semantic model that comprehensively describes spatial correlation, time series relationship and business logic constraints.

7. The data management method for power space data according to claim 1 is characterized in that: The constructed semantic model is used to extract features of power space data through embedding and attention mechanisms to generate multi-level semantic features, including: Embed the spatial dimension of power spatial data to transform the geographical location information, connection relationship and physical topology characteristics of the equipment into high-dimensional feature representation that can be processed by the model; The time dimension of power spatial data is serialized and embedded, and the historical records and time dependencies of equipment operation status are mapped into continuous time feature representation to capture the impact of key time points; The business logic dimension of power spatial data is processed by rule embedding, and the logical dependency relationship between power business rules and events is converted into a logical feature vector, retaining the causal relationship between events; The above embedded features are weighted through the attention mechanism, and the weight of each feature is dynamically assigned. Key features are extracted according to the importance of spatial correlation, time series patterns and business rules, generating multi-level semantic features including spatial correlation, time series relations and business logic constraints.

8. The data management method for electric power spatial data according to claim 1, characterized in that: The method of constructing a semantic graph by using the extracted multi-level semantic features through graph database technology includes: The extracted multi-level semantic features are divided into feature sets of nodes and edges, where node features include the spatial location, time state and business logic attributes of power equipment, and edge features include the physical connection relationship between devices, the time dependency relationship between events, and the logical rule constraint relationship; The graph structure is defined based on node features and edge features, where nodes represent power equipment or key events, and edges represent associations between equipment or logical dependencies between events; Create a storage structure for nodes and edges in the graph database, and attach attributes to each node and edge, including spatial characteristics, temporal characteristics, and contextual information of business rules; According to the spatiotemporal changes of power spatial data, the nodes and edges in the graph structure are dynamically updated to ensure that the semantic graph can reflect changes in equipment status and adjustments to business logic in real time.

9. The data management method for electric power spatial data according to claim 1, characterized in that: According to the preliminary repair suggestion, according to the logical rules generated by the semantic model, the repair content of the abnormal data, redundant data and conflicting data is subjected to semantic consistency verification, including: Extracting the target data and its context information involved in the preliminary repair suggestion, including the source of the repair data, applicable logical rules and contextual association conditions; Based on the logical rule base generated by the semantic model, the repair content is checked one by one to verify whether the repair data meets the physical constraints of the equipment, time series logic and business operation specifications; For repair content involving multiple data sources or relying on contextual conditions, the logical reasoning module verifies whether the repaired data is consistent with the associated data, including the logical integrity of the event chain and the logical continuity of the device status.

10. A data management system for power space data, characterized in that: include: A modeling unit is used to jointly model the spatial dimension, time dimension and business logic dimension of the power spatial data based on the spatiotemporal characteristics of the power spatial data and the power business logic, using a deep learning model to generate a semantic model of the spatiotemporal characteristics and business logic constraints; A generation unit, used to extract features of the power space data by using the constructed semantic model through embedding and attention mechanisms to generate multi-level semantic features, wherein the multi-level semantic features include spatial correlation, time series relationship, and business logic constraints of the power space data; A construction unit is used to construct a semantic graph by using the extracted multi-level semantic features through graph database technology; based on the semantic graph, a global power space data knowledge base is generated, and logical associations and contextual relationships between power space data are dynamically marked in the knowledge base; A detection unit, configured to detect abnormal data, redundant data and conflicting data in the power space data based on the global power space data knowledge base and in combination with a data cleaning algorithm, and generate preliminary repair suggestions; A verification unit, configured to perform semantic consistency verification on the repair contents of abnormal data, redundant data and conflicting data according to the preliminary repair suggestion and the logic rules generated by the semantic model; A repair unit is used to repair abnormal data, redundant data and conflicting data according to the verification result of the semantic consistency check to generate a managed power spatial data set.

Images