Power grid service package configuration method and system combined with grid analysis

Abstract

The present invention discloses a method and system for configuring a power grid service package combined with grid analysis, and relates to the field of smart grid technology. The method includes: parsing the content of a power grid service package to obtain a service configuration dimension; determining the grid granularity of the configuration dimension; performing grid segmentation on each service configuration dimension to determine the segmented grid; analyzing each grid unit according to the segmented grid to extract power characteristics; performing feature aggregation on the power characteristics to obtain grid aggregation characteristics; and performing feature projection according to the mapping relationship between the service configuration dimension and the grid aggregation characteristics to configure the power grid service package. The present invention solves the technical problems in the prior art that the configuration of power grid service packages lacks precise adaptation, resource allocation is inefficient, and spatial structure characteristics and correlation relationship mining are not fully utilized, and achieves the technical effect of accurately adapting the service package configuration to the power grid operation needs, improving the pertinence and efficiency of power grid services, and enhancing the reliability of power grid operation.

Description

Technical Field

[0001] The present invention relates to the field of smart grid technology, and in particular to a method and system for configuring a grid service package in combination with grid analysis. Background Art

[0002] In the field of power grid services, with the increasing diversification of user needs and the continuous expansion of the scale of power grids, the traditional configuration method of power grid service packages faces many challenges. Traditional methods often lack a comprehensive and in-depth content analysis of power grid service packages, and cannot accurately determine the service configuration dimensions, resulting in a disconnect between service configuration and actual power grid operation needs. For example, it is difficult to design service package content in a targeted manner when faced with differences in power consumption characteristics in different regions, diverse service expectations of user groups, and complex and changeable power facility conditions. At the same time, when determining the granularity of the configuration dimension, there is a lack of scientific and effective means, and it relies more on experience or simple division methods, and fails to fully consider the characteristics of service parameters, making the grid division unreasonable. This may cause the grids in some areas to be too large and unable to reflect the changes in local power characteristics in detail; while the grids in some areas are too small, increasing the computational burden and possibly ignoring the correlation between regions. In addition, when analyzing power grid data, traditional technologies fail to effectively integrate multi-source information such as user power consumption portrait characteristics, power grid service fault records, and user service feedback needs, and cannot accurately extract comprehensive power characteristics, nor can they mine the inherent correlation between these characteristics.

[0003] The existing technology has technical problems such as lack of precise adaptation of power grid service package configuration, inefficient resource allocation, under-utilization of spatial structure characteristics and insufficient mining of correlation relationships. Summary of the invention

[0004] The present application provides a method and system for configuring a power grid service package in combination with grid analysis, which is used to solve the technical problems in the prior art of lack of precise adaptation of power grid service package configuration, inefficient resource allocation, underutilization of spatial structure characteristics and insufficient mining of association relationships.

[0005] In view of the above problems, the present application provides a method and system for configuring a power grid service package combined with grid analysis.

[0006] In a first aspect of the present application, a method for configuring a power grid service package in combination with grid analysis is provided, the method comprising:

[0007] The power grid service package is content parsed to obtain a service configuration dimension; based on the service parameter influence characteristics of the service configuration dimension, the grid granularity of the configuration dimension is determined; according to the grid granularity of the configuration dimension, each service configuration dimension is grid-segmented to determine the segmented grid; according to the segmented grid, each grid unit is analyzed to extract power characteristics, and the power characteristics include one or more of power demand, fault characteristics, power facility status, and user demand; based on the grid granularity of the configuration dimension, the power characteristics are feature aggregated to obtain grid aggregate characteristics; according to the mapping relationship between the service configuration dimension and the grid aggregate characteristics, feature projection is performed to configure the power grid service package.

[0008] A second aspect of the present application provides a grid service package configuration system combined with grid analysis, the system comprising:

[0009] A service configuration dimension acquisition module, the service configuration dimension acquisition module is used to perform content parsing on the power grid service package to obtain the service configuration dimension; a grid granularity determination module, the grid granularity determination module determines the grid granularity of the configuration dimension based on the service parameter impact characteristics of the service configuration dimension; a segmentation grid determination module, the segmentation grid determination module is used to perform grid segmentation on each service configuration dimension according to the grid granularity of the configuration dimension to determine the segmentation grid; a power feature extraction module, the power feature extraction module is used to analyze each grid unit according to the segmentation grid to extract power features, the power features including one or more of power demand, fault features, power facility status, and user demand; a grid aggregation feature acquisition module, the grid aggregation feature acquisition module performs feature aggregation on the power features based on the grid granularity of the configuration dimension to obtain grid aggregation features; a power grid service package configuration module, the power grid service package configuration module is used to perform feature projection and configure the power grid service package according to the mapping relationship between the service configuration dimension and the grid aggregation feature.

[0010] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0011] The content of the power grid service package is parsed to obtain the service configuration dimension; based on the service parameter influence characteristics of the service configuration dimension, the grid granularity of the configuration dimension is determined; according to the grid granularity of the configuration dimension, each service configuration dimension is grid-segmented to determine the segmented grid; according to the segmented grid, each grid unit is analyzed to extract the power characteristics; based on the grid granularity of the configuration dimension, the power characteristics are feature aggregated to obtain the grid aggregation characteristics; according to the mapping relationship between the service configuration dimension and the grid aggregation characteristics, feature projection is performed to configure the power grid service package. The service package configuration that accurately adapts to the power grid operation needs is achieved, the pertinence and efficiency of the power grid service are improved, and the reliability of the power grid operation is enhanced. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0013] Figure 1 A schematic flow chart of a method for configuring a power grid service package in combination with grid analysis provided in an embodiment of the present application.

[0014] Figure 2 A schematic diagram of the structure of a power grid service package configuration system combined with grid analysis provided in an embodiment of the present application.

[0015] Explanation of reference numerals: service configuration dimension acquisition module 10 , grid granularity determination module 20 , segmented grid determination module 30 , power feature extraction module 40 , grid aggregation feature acquisition module 50 , power grid service package configuration module 60 . DETAILED DESCRIPTION

[0016] The present application provides a method and system for configuring a power grid service package in combination with grid analysis, so as to solve the technical problems in the prior art of lack of precise adaptation of power grid service package configuration, inefficient resource allocation, underutilization of spatial structure characteristics and insufficient mining of association relationships.

[0017] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0018] Embodiment 1, as Figure 1As shown, the present application provides a method for configuring a power grid service package in combination with grid analysis, the method comprising:

[0019] Step S100: parse the content of the power grid service package to obtain service configuration dimensions.

[0020] Specifically, the user electricity consumption logs in the target area are collected to obtain the user electricity consumption image features, such as electricity consumption habits, peak hours, etc. At the same time, the grid service fault records and user service feedback requirements are obtained, and the user grid service requirements are extracted based on these fault records and feedback requirements, such as the demand for power supply stability and the demand for timely fault repair. Then, with the preset type of the grid service package as the target, the user electricity consumption image features and user grid service requirements are preprocessed and input into the association mining model. The model divides the input information recognition into explicit feature information and implicit feature information through the encoding module. The multi-order iterative association layer in the information association module (at least including the first-order association layer and the second-order association layer, the first-order association layer extracts the feature record location according to the explicit feature information to perform feature association mining to obtain the first-order mining feature, the second-order association layer obtains the time series feature according to the feature record location, and combines the first-order mining feature and the implicit feature information to perform feature association mining) performs association mining on the input feature information, and the reply generation module performs correlation calculation and outputs the association influence according to the feature description of the association mining. In this way, service configuration dimensions such as service coverage, service response time, etc. are obtained according to the associated influence with the preset type of the power grid service package.

[0021] Step S200: Determine the grid granularity of the configuration dimension based on the service parameter impact characteristics of the service configuration dimension.

[0022] Specifically, the service configuration dimensions that have been obtained are deeply analyzed to clarify the influencing characteristics of various service parameters involved in the operation of the power grid. These characteristics cover many aspects, such as the potential impact of different configuration dimensions on power transmission efficiency, power supply reliability, and cost control. Through detailed analysis, the intrinsic relationship between service configuration dimensions and key indicators of power grid operation is comprehensively sorted out. Then, a suitable clustering algorithm, such as the K-means clustering algorithm, is used to cluster these service parameter influencing characteristics, and service configuration dimensions with similar influencing characteristics are classified into one category to obtain clustering results. In this process, it is necessary to reasonably select the parameters of the clustering algorithm according to the characteristics of the actual power grid data and the analysis requirements to ensure the accuracy and effectiveness of the clustering results. Finally, the cluster center of the clustering result is used as the center position of the grid unit, and the size of the grid unit is determined according to the clustering density presented by the clustering result. If the density of a cluster is large, it means that the correlation of this part of the service configuration dimension is strong, and a smaller grid unit can be set accordingly to achieve a more detailed analysis; on the contrary, for areas with a smaller clustering density, a larger grid unit can be set to ensure the effectiveness of the analysis while improving the calculation efficiency. In this way, the grid granularity suitable for each service configuration dimension is determined, laying a solid foundation for subsequent grid segmentation and precise analysis.

[0023] Step S300: grid segmenting each service configuration dimension according to the grid granularity of the configuration dimension to determine the segmented grid.

[0024] Specifically, according to the grid granularity of the determined configuration dimension, precise grid segmentation operations are carried out for each service configuration dimension. For each service configuration dimension, such as the dimension of geographic area coverage and the dimension of user group classification, detailed divisions are performed according to the established grid granularity. Taking the geographic area coverage as an example, if the grid granularity is set to the area size in square kilometers, the entire service area is divided into multiple regular or irregular grid units according to the unit, and these grid units together constitute the segmentation grid for the service configuration dimension. In the division process, the characteristics of the service configuration dimension and the actual power grid operation requirements must be fully considered to ensure that the grid segmentation can accurately reflect the differences between different regions or categories, and maintain a certain logical coherence and data integrity, thereby providing a reasonable and effective structural framework for the subsequent in-depth analysis of each grid unit, so that the configuration of the power grid service package can be scientifically planned and optimized at the fine grid level.

[0025] Step S400: Analyze each grid unit according to the segmented grid to extract power characteristics, where the power characteristics include one or more of power demand, fault characteristics, power facility status, and user demand.

[0026] Specifically, based on the segmented grids determined in the previous steps, a comprehensive and in-depth analysis is conducted for each grid unit to extract power characteristics covering many aspects. For power demand characteristics, by collecting the power consumption and power consumption data of various users (including residents, commercial users, industrial users, etc.) in the grid unit in different time periods (such as working days, holidays, peak and valley periods, etc.), data analysis technology (data mining algorithm) is used to accurately grasp the power demand pattern and change law in the grid unit, including the total amount of power demand, peak-valley difference, growth trend, etc. In terms of fault characteristics, the fault records related to the grid unit in the historical data of power grid operation are deeply mined, and the frequency, type (such as short circuit fault, open circuit fault, overvoltage fault, etc.), occurrence time distribution and fault duration of fault are counted in detail. Through the systematic analysis of these fault data, the weak links and high-incidence areas of the power grid system in the grid unit are identified. For the status of power facilities, sensor technology (such as temperature sensors, pressure sensors, current sensors, etc.) is used to monitor the operating parameters of various power facilities (such as substations, transmission lines, distribution transformers, etc.) in the grid unit in real time, including equipment temperature, operating voltage, current load, etc., and combined with equipment maintenance records, operating years and other information, the health of power facilities is comprehensively evaluated to determine whether there are potential operating risks or performance degradation problems. In addition, through user feedback channels (such as opinions and suggestions, questionnaires, etc.), the demand information of users in the grid unit is widely collected, including expectations for power supply quality (such as voltage stability, power supply reliability, etc.), sensitivity to electricity prices, and willingness to demand new power services (such as smart home power management, distributed energy access, etc.), etc. These user demand information is sorted and summarized and included in the analysis scope as important power characteristics. Through the comprehensive extraction of the above power demand, fault characteristics, power facility status and user needs, rich and detailed data support is provided for the subsequent precise configuration of power grid service packages.

[0027] Step S500: Based on the grid granularity of the configuration dimension, the power features are aggregated to obtain grid aggregated features.

[0028] Specifically, based on the grid granularity of the configuration dimension, an effective feature aggregation operation is performed on the power features extracted from each grid unit. First, a reasonable adjacency granularity range is set according to the grid granularity. This range determines the range of surrounding grid units considered when performing feature aggregation. Next, each grid unit is regarded as a node, and the connection edges are set according to the association between grid units in terms of geographic space, power supply relationship, etc., so as to construct the power grid service data graph structure. Then, the neighborhood of the power grid service data graph structure is extracted according to the set adjacency granularity range to form a neighborhood aggregation circle. In this neighborhood aggregation circle, the connection edge relationship is used as a link, and each grid unit is used as the aggregation center. According to the position of the aggregation center in the neighborhood aggregation circle and the distance relationship of the connection edge, the aggregation direction and aggregation path are carefully set. For example, starting from the grid unit farthest from the aggregation center, it gradually advances to the aggregation center along the set aggregation path. In the feature aggregation process, the features of the grid cells are weighted averaged from far to near from the aggregation center. A relatively small weight is given to the grid cells far from the aggregation center, while a relatively large weight is given to the grid cells close to the aggregation center. This can highlight the characteristic influence of the central grid cell and its surrounding closely related grid cells. In this way, the power features of each grid cell in the neighborhood aggregation circle are organically integrated, and finally the grid aggregation features of each grid cell are obtained. These features integrate the power information of the grid cell and the surrounding related grid cells, and can more comprehensively and accurately reflect the overall operation status and characteristics of the power grid in the region.

[0029] Step S600: According to the mapping relationship between the service configuration dimension and the grid aggregation feature, feature projection is performed to configure a power grid service package.

[0030] Specifically, based on the intrinsic mapping relationship between the determined service configuration dimensions and grid aggregation characteristics, key feature projection operations are carried out to achieve the reasonable configuration of the power grid service package. First, the logical connection between the service configuration dimensions and the grid aggregation characteristics is deeply analyzed. This connection is established through deep mining and analysis of the power grid operation data. It clarifies how different grid aggregation characteristics correspond to and affect each service configuration dimension. Then, based on this mapping relationship, the rich information contained in the grid aggregation characteristics is projected onto the corresponding service configuration dimensions. For example, if the grid aggregation characteristics of a grid unit show that the power demand is growing rapidly and there is a certain aging risk in the status of the power facilities, then in the service configuration dimension, the power supply strategy will be adjusted accordingly, such as increasing the power generation capacity, optimizing the grid structure, and accelerating the maintenance and upgrade plan of the power facilities; if the grid aggregation characteristics reflect that users are highly concerned about the reliability of power supply and the frequency of faults in the area is high, then in the service configuration, emphasis will be placed on strengthening the fault monitoring and repair capabilities, improving the emergency response speed, and increasing the backup power supply configuration. Through this precise feature projection, the various service contents in the power grid service package are comprehensively optimized, including but not limited to power supply plans, fault handling mechanisms, user service types and quality standards, to ensure that power grid services can closely meet the actual needs of each grid unit, achieve efficient allocation and utilization of power grid resources, improve the operating efficiency, reliability and user satisfaction of the entire power grid system, and provide strong support for the sustainable development of the power grid.

[0031] In a possible implementation, step S100 further includes:

[0032] Step S110: Collect electricity usage logs of users in the target area to obtain the user's electricity usage profile features.

[0033] Step S120: Obtaining power grid service fault records and user service feedback requirements.

[0034] Step S130: extracting user grid service demands according to the grid service fault records and user service feedback demands.

[0035] Step S140: Taking the preset type of the grid service package as the target, performing correlation impact analysis according to the user's electricity consumption profile characteristics and the user's grid service demand, and obtaining the service configuration dimension according to the correlation impact with the preset type of the grid service package.

[0036] Specifically, electricity consumption log data of various users in the target area are widely collected. These data cover electricity consumption information of different groups such as residential users, commercial users and industrial users over a long period of time. Through in-depth mining of massive electricity consumption data, data processing technology and algorithms are used to extract portrait features that can characterize the user's electricity consumption behavior pattern. For example, by analyzing the user's daily electricity consumption fluctuation curve, its peak and trough periods can be determined, thereby judging the user's daily activity patterns; according to the changes in electricity consumption in different seasons, the user's seasonal electricity demand characteristics can be understood; at the same time, the analysis of data such as power factor and power consumption frequency helps to grasp the type and usage of users' electrical equipment, and provide basic data support for the subsequent precise configuration of power grid services.

[0037] At the same time, detailed grid service fault records are obtained from the grid operation and management system, including key information such as the time, location, type (such as short circuit, circuit break, overvoltage, etc.) of the fault, and duration of the fault. These fault records reflect the problems and weak links in the actual operation of the grid. In addition, user service feedback needs are collected through various channels, such as online questionnaires and community opinion collection, and users' opinions and suggestions on power supply quality, service response speed, and reasonable electricity charges are widely listened to, so as to fully understand users' expectations and dissatisfaction with grid services, and provide direction for further optimizing grid services.

[0038] Based on the acquired grid service fault records and user service feedback needs, the user grid service needs contained therein are deeply analyzed. For example, if a certain area frequently experiences voltage fluctuation faults and users report that their electrical equipment is greatly affected, then it can be extracted that users in this area have a strong demand for stable power supply voltage; if users generally report that their electricity bills are unclear and the payment methods are inconvenient, this indicates that users have a need to improve electricity bill management and payment services. Through a comprehensive analysis of these data, the actual needs of users are converted into specific grid service demand indicators, such as power supply reliability improvement requirements, service convenience improvement goals, etc., so that they can be met in a targeted manner in the configuration of the grid service package.

[0039] The preset type of power grid service package is set as the core goal, and this is used as a guide to conduct an in-depth correlation and influence analysis on the portrait characteristics of user electricity consumption and user power grid service needs. First, the portrait characteristics of user electricity consumption are sorted out in detail, including user electricity consumption habits (such as the distribution of daily peak and valley hours), power load characteristics (such as average power consumption, maximum power consumption) and power consumption behavior patterns (such as whether there is a periodic power consumption pattern) and other information. At the same time, a comprehensive analysis of user power grid service needs is carried out to clarify the specific needs of users in terms of power supply quality (such as voltage stability, frequency deviation requirements), service response timeliness (such as the expected repair time after fault reporting), and electricity price rationality (such as expectations for electricity price pricing methods). Then, scientific data analysis methods and models are used to explore the inherent connection between the portrait characteristics of user electricity consumption and user power grid service needs. For example, through analysis, it is found that the peak power consumption of a certain type of commercial user is concentrated in a specific time period and the power load is large (portrait characteristics), and its demand for power supply reliability and emergency power support is also stronger (service demand), and there is a significant correlation between the two. Finally, the service configuration dimensions are determined based on the degree of correlation between these characteristics and the demand and the preset type of the grid service package. If a certain characteristic or demand is highly correlated with a preset type of grid service package, for example, the situation of the above-mentioned commercial users is closely related to the high-reliability power supply service package type, then power supply reliability, emergency power supply configuration, etc. can be determined as key service configuration dimensions. In this way, starting from the actual situation of user electricity consumption and combining the preset service package type, the dimensions that have an important impact on the grid service configuration are accurately located, providing a scientific basis for the reasonable configuration of subsequent grid service packages.

[0040] In a possible implementation, step S140 further includes:

[0041] Step S141: pre-process the user's electricity consumption portrait features and the user's grid service requirements, input them into the association mining model, and obtain matching side demand association features.

[0042] Step S142: pre-processing the preset type of the power grid service package, inputting it into the association mining model, and obtaining target-side service association features.

[0043] Step S143: input the matching-side demand-related features and the target-side service-related features into the association mining model to obtain association influence.

[0044] Specifically, preprocessing operations are performed on the collected user power consumption portrait features and user grid service needs. In terms of portrait features, the massive power consumption data is cleaned to remove abnormal values ​​and erroneous data, such as correcting erroneous readings caused by meter failures and filling missing values ​​caused by data transmission interruptions. At the same time, the data is standardized to make different types of power consumption data (such as power consumption, power consumption, etc.) comparable and converted into a format suitable for model input, such as normalizing the power consumption data so that its value is within a specific range. For user grid service needs, they are classified and sorted, and the user's needs in terms of power supply quality, service response, etc. are refined and quantified, such as quantifying the power supply quality demand according to specific indicators such as voltage stability and frequency deviation. Then, the preprocessed user power consumption portrait features and user grid service needs are input into the association mining model. The model uses internal complex algorithms and calculation logic to deeply explore the potential correlation between these data, thereby obtaining matching side demand correlation features, which can accurately reflect the inherent connection between user power consumption behavior and demand, such as discovering strong correlation patterns between certain specific power consumption habits and specific service needs.

[0045] Preprocess the preset types of power grid service packages, sort out and encode the service content, service objectives, service scope and other information of the preset types, and convert them into a form that the model can recognize and process, such as classifying and encoding the service types and quantifying the service objectives. Then input the preprocessed preset types of power grid service packages into the association mining model. The model analyzes the association between the preset types and the various elements of power grid operation according to its internal rules and algorithms, so as to obtain the target side service association characteristics. These characteristics can reflect the matching relationship between different preset types of service packages and power grid operation requirements, such as determining the degree of association between a certain preset service package type and specific power grid operation parameters (such as power supply reliability indicators, power transmission efficiency, etc.).

[0046] Finally, the matching side demand association features and target side service association features obtained previously are input into the association mining model again. Based on its powerful data analysis and computing capabilities, the model conducts a comprehensive comparison and in-depth analysis of these two types of association features, calculates the degree of mutual influence between them, and thus obtains the association influence. For example, through calculation, it is found that there is a high degree of mutual influence between the association features of a certain type of user demand and the target side service association features of a preset type of a specific service package, which means that this type of user demand has an important influence on the configuration of this preset type of service package, thereby providing a key basis for determining the service configuration dimension, helping to accurately locate the service content and configuration parameters that should be focused on and optimized in the power grid service package, so as to better meet user needs and improve the efficiency of power grid operation.

[0047] In a possible implementation, step S141 further includes:

[0048] Step S1411: construct an association mining model architecture, including an encoding module, an information association module, and a response generation module, wherein the encoding module is used to identify and divide the input information into explicit feature information and implicit feature information, the information association module includes a multi-order iterative association layer to perform association mining on the input feature information, and the response generation module performs correlation calculation and outputs an association result based on the feature description of the association mining; wherein the multi-order iterative association layer includes at least a first-order association layer and a second-order association layer, the first-order association layer extracts the feature according to the explicit feature information, performs feature association mining at the record position, and obtains the first-order mining feature, the second-order association layer obtains the time series feature according to the record position of the feature, and performs feature association mining in combination with the first-order mining feature and the implicit feature information.

[0049] Specifically, an association mining model architecture of IAMM (Intelligent Analysis and Management Mining) was constructed, which consists of three important modules. The modules work together to achieve deep mining and association analysis of complex information, thereby providing strong support for the configuration of power grid service packages.

[0050] The encoding module is the data preprocessing hub of the model. When input information such as the user's electricity consumption profile features and the user's grid service needs enter the model, the encoding module uses a series of advanced technologies to perform identification and segmentation tasks. For the user's electricity consumption profile features, it may contain many data dimensions, such as electricity consumption, power consumption, power factor, etc. in different time periods, as well as information related to electricity consumption behavior, such as whether electricity is concentrated in a specific time period. The encoding module uses text classification technology and feature extraction algorithms in natural language processing (such as rule-based classifiers and principal component analysis PCA algorithms) to subdivide these input information into explicit feature information and implicit feature information. Explicit feature information is data that can be directly observed on the surface, such as specific electricity consumption values, clear electricity consumption time periods, etc.; implicit feature information is hidden behind the data and needs to be revealed through further data mining and pattern recognition technology, such as discovering potential electricity consumption habits and preferences in users' long-term electricity consumption behavior patterns through cluster analysis, and identifying electricity consumption fluctuation patterns that may imply potential needs for power supply stability through anomaly detection algorithms.

[0051] The information association module is the core processing unit of the model, and its internal multi-order iterative association layer undertakes the key association mining task. The first-order association layer focuses on the explicit feature information, and uses the feature association algorithm based on the position index (such as the position-sensitive hashing algorithm) to establish the preliminary association relationship between the features according to the position information of the explicit feature information in the data record, thereby obtaining the first-order mining features. For example, the position of the electricity consumption data in the time series record is associated with the position of the power consumption data in the same period, and the possible correlation patterns between the two are mined, such as the specific relationship between electricity consumption and power consumption during the peak period of electricity consumption. The second-order association layer further deepens the mining work. It first uses time series analysis technology (such as wavelet analysis) to obtain time series features according to the record location of the feature. These time series features can accurately capture the trend of data changes over time, periodic laws, and seasonal fluctuations, such as the change pattern of user electricity consumption between weekdays and weekends and different seasons. Then, the second-order association layer combines the first-order mining features and implicit feature information, and uses the association mining algorithm that integrates the deep learning architecture (such as the long short-term memory network LSTM) to fully consider the time series relationship and potential patterns of the features, and mine more in-depth and comprehensive association features. For example, the first-order mining of the relationship between electricity consumption and power during peak hours is combined with the user's electricity consumption habits and preferences obtained through cluster analysis in the implicit features, and the seasonal electricity consumption fluctuations in the time series features, revealing a more complex relationship between electricity consumption and power during peak hours under specific seasons and specific electricity consumption habits, providing rich and in-depth information for a comprehensive understanding of user electricity consumption behaviors and needs.

[0052] The reply generation module calculates the correlation based on the complex feature description mined by the information association module, and finally outputs the correlation result. This module uses typical correlation analysis in machine learning to quantitatively evaluate the degree of correlation between different types of features. For example, the relationship between the user's power consumption portrait characteristics and the preset type of the power grid service package is determined by calculating the typical correlation coefficient between the two to determine the degree of correlation between them. This correlation result is output in the form of clear and definite numerical or logical relationships, which provides a direct and reliable data basis for subsequent work such as determining the service configuration dimension, enabling decision makers to accurately grasp the inherent correlation between various factors, so as to formulate a scientific and reasonable power grid service package configuration strategy that meets actual needs, realize effective feature division and deep correlation mining of input information, and enhance the model's ability to analyze and discover associations for complex power grid service data.

[0053] In a possible implementation, step S200 further includes:

[0054] Step S210: performing configuration impact parameter analysis on the service configuration dimension to obtain service parameter impact characteristics.

[0055] Step S220: performing cluster analysis on the service parameter impact characteristics by using a clustering algorithm to obtain a clustering result.

[0056] Step S230: taking the cluster center of the clustering result as the grid unit center, determining the size of the grid unit according to the clustering density of the clustering result, and obtaining the grid granularity of the configuration dimension.

[0057] Specifically, an in-depth analysis of the configuration influencing parameters is carried out for the determined service configuration dimensions. The service configuration dimensions cover many aspects, such as geographical coverage, user group type, power supply reliability requirements, power service response time, etc. For the geographical coverage dimension, its configuration influencing parameters may include the topography in the region (such as mountainous areas and plain areas have different impacts on power grid construction and power transmission), population density (there are differences in electricity demand and power facility layout requirements between densely populated areas and sparsely populated areas), economic development level (different electricity load characteristics and expectations for power service quality in developed and underdeveloped areas), etc. Through a detailed analysis of these parameters, the characteristics that can accurately characterize the impact of each service configuration dimension on power grid operation and service provision are extracted, namely, service parameter impact characteristics. For example, the service parameter impact characteristics analyzed from the geographical coverage dimension may include regional power demand density, power transmission difficulty coefficient, etc. These characteristics can intuitively reflect the key influencing factors of this dimension on power grid service configuration.

[0058] Use a suitable clustering algorithm to perform cluster analysis on the service parameter influencing features. When using the K-means clustering algorithm, firstly, it is necessary to preliminarily determine the number of clusters K based on experience or data analysis, and then randomly select K initial cluster centers. Treat each service parameter influencing feature as a data point, calculate its distance from each cluster center (the distance measurement method can be selected according to the data characteristics, such as Euclidean distance, cosine distance, etc.), and assign it to the cluster to which the nearest cluster center belongs. After the assignment is completed, recalculate the new cluster center of each cluster (usually the mean of the data points in the cluster), and repeat the above process of assigning and updating the cluster center until the convergence condition is met (such as the cluster center no longer changes significantly or reaches the maximum number of iterations). Through such a clustering process, similar service parameter influencing features are clustered together to form different clusters, thereby obtaining clustering results. For example, clustering features such as power demand density and power transmission difficulty coefficient related to the coverage of the geographical area may result in different cluster clusters, representing different types of regional feature combinations.

[0059] Finally, the grid granularity of the configuration dimension is determined based on the clustering results, and the cluster center in the clustering results is used as the center position of the grid unit. This is because the cluster center represents the typical value or average level of the service parameter impact characteristics in the cluster. Then, the size of the grid unit is determined according to the clustering density of the clustering results. The clustering density reflects the density of data points in the cluster. If the clustering density of a cluster is large, it means that the service parameter impact characteristics in the cluster are relatively similar and concentrated. At this time, a relatively small grid unit can be set to more finely analyze and process the power grid service configuration in the area; on the contrary, if the clustering density is small, it means that the characteristics in the cluster are large or sparsely distributed, then a larger grid unit can be set to ensure the effectiveness of the analysis while improving the calculation efficiency. For example, for areas with a large clustering density of power demand density and power transmission difficulty coefficient characteristics, it means that these characteristics in the area are relatively stable and similar, and a smaller grid unit can be set for more detailed power grid service planning; while for areas with a small clustering density, a larger grid unit can be used to reduce the calculation complexity and still grasp the overall characteristics. In this way, the grid granularity suitable for each service configuration dimension is determined, providing a reasonable basic framework for subsequent grid segmentation and precise analysis, ensuring that the grid division conforms to the characteristics of the service configuration dimension and improving the scientific nature of the grid segmentation.

[0060] In a possible implementation, step S500 further includes:

[0061] Step S510: setting an adjacent granularity range according to the grid granularity of the configuration dimension.

[0062] Step S520: Using grid units as nodes, setting connection edges according to association relationships between grid units, and constructing a power grid service data graph structure.

[0063] Step S530: performing neighborhood extraction on the power grid service data graph structure according to the adjacency granularity range to obtain a neighborhood aggregation circle.

[0064] Step S540: performing node feature aggregation according to the neighborhood aggregation circle and the connection edge relationship to obtain the grid aggregation feature.

[0065] Specifically, the adjacent granularity range is determined according to the grid granularity of the configuration dimension. The grid granularity determines the range of the service configuration space represented by each grid unit. On this basis, the adjacent granularity range is used to define the range of adjacent grid units considered when performing feature aggregation. For example, if the grid granularity is a grid unit divided by a specific geographical area, the adjacent granularity range can be set to one or more layers of grid units directly adjacent to the grid unit. The setting of this range needs to comprehensively consider the characteristics and analysis requirements of the power grid service data, ensuring that sufficient spatial dependencies can be captured, but not too broad to cause excessive calculations or introduce too much irrelevant information. If the adjacent range is too small, the impact of some important adjacent areas on the characteristics of the current grid unit may be ignored; if it is too large, the aggregated features may be interfered by irrelevant factors from too far away areas.

[0066] Each grid unit is regarded as a node in the graph, and the connection edges are set according to the association relationship between the grid units, so as to construct the graph structure of the power grid service data. The association relationship between grid units can be determined based on a variety of factors, such as geographical proximity (there is a strong power transmission and service association between physically adjacent grid units), the direction and intensity of power flow (power flows from one grid unit to another to establish a connection), the similarity of user groups (grid units with similar power demand and behavior patterns are connected), etc. The connection edges constructed through these association relationships connect the nodes of each grid unit to form a graph structure that can reflect the spatial structure and dependency relationship of power grid service data. For example, in a city power grid, there may be a strong connection relationship between grid units located on the same power supply line or grid units in adjacent commercial areas. These relationships are reflected in the graph structure through connection edges, providing a structural basis for subsequent feature aggregation.

[0067] According to the set adjacency granularity range, the constructed power grid service data graph structure is subjected to neighborhood extraction operation. Starting from each grid unit node, the neighboring nodes around it are determined according to the adjacency granularity range. These neighboring nodes and their connecting edges with the central node constitute the neighborhood aggregation circle. The extraction of the neighborhood aggregation circle is to focus on the area with direct or indirect influence around each grid unit, so as to more effectively capture the local spatial dependency in the subsequent aggregation process. For example, for a central grid unit, if the adjacency granularity range is set to a directly adjacent layer of grid units, then its neighborhood aggregation circle is composed of those directly adjacent grid units and their connecting edges; if the adjacency granularity range is expanded to two layers of adjacent grid units, the neighborhood aggregation circle includes all grid units and connecting edges within the two layers, so that the impact of the surrounding environment on the characteristics of the central grid unit can be more comprehensively considered.

[0068] Finally, node feature aggregation operations are performed according to the neighborhood aggregation circle and the connection edge relationship to obtain grid aggregation features. In the neighborhood aggregation circle, each node (grid unit) has its own power characteristics, such as power demand, fault characteristics, power facility status, etc. Based on the aggregation rules of graph neural networks, aggregation is achieved by passing and fusing node feature information along the connection edges. For example, using a message passing mechanism, each node passes its own feature information to adjacent nodes along the connection edges, while receiving information from adjacent nodes, and fusing this information according to certain aggregation functions (such as summation, average, weighted summation, etc.). For neighbor nodes that are closer to the central node, a larger weight can be given because their impact on the central node's characteristics is usually more direct and important; while the weights of nodes farther away are relatively small. In this way, the characteristics of all nodes in the neighborhood aggregation circle are integrated to obtain the grid aggregation characteristics of each grid unit. These aggregation characteristics not only contain the information of the grid unit itself, but also integrate the influence of the surrounding related grid units, thereby more comprehensively and accurately reflecting the overall operation status and structural characteristics of the power grid in the region. The graph neural network aggregation logic is used to construct the power grid service data graph structure and perform feature aggregation, effectively capturing the spatial dependency and structural characteristics of the power grid service data.

[0069] In a possible implementation, step S540 further includes:

[0070] Step S541: Taking each grid unit as an aggregation center, according to the aggregation center in the neighborhood aggregation circle, and according to the distance relationship of the connecting edges, an aggregation direction and an aggregation path are set.

[0071] Step S542: According to the aggregation direction and the aggregation path, weighted average of the features of the grid units is performed in sequence from far to near from the aggregation center to obtain the grid aggregation features of the grid units.

[0072] Specifically, for each grid unit, take it as the aggregation center, and in the constructed neighborhood aggregation circle, carefully set the aggregation direction and aggregation path according to the distance relationship of the connecting edges. The distance relationship of the connecting edges reflects the degree of spatial or logical association between grid units. For grid units that are closer to the aggregation center, the distance of their connecting edges is relatively short, which means that they are more correlated with the aggregation center and have a more direct impact on the characteristics of the aggregation center. Therefore, when setting the aggregation direction, priority is given to aggregating information from these neighboring units. For example, if a distance-based breadth-first search strategy is adopted, the aggregation direction will gradually expand outward from the neighboring units directly connected to the aggregation center. The aggregation path is the information transmission route formed by these connecting edges, which determines the specific way for information to flow from one grid unit to another. For example, in a simple two-dimensional grid structure, if the aggregation center is located in the center, its aggregation path may first transmit information to the four adjacent grid units along the horizontal and vertical connecting edges, and then expand to the eight grid units with a distance of two grids in turn, and so on, forming an information transmission network from the inside to the outside, gradually covering the neighborhood aggregation circle.

[0073] According to the set aggregation direction and aggregation path, the weighted average operation of the grid unit features is performed in order from far to near from the aggregation center. For each grid unit, its own power characteristics (such as power demand, fault characteristics, power facility status, etc.) have different importance weights in the aggregation process. Although the grid unit far from the aggregation center also has a certain influence on the characteristics of the central unit, it is relatively weak, so it is given a smaller weight; while the grid unit closer to the aggregation center has a greater influence on the central unit, and the weight is correspondingly larger. For example, the distance decay function is used to determine the weight, such as the weight is proportional to the inverse of the distance, the weight of the grid unit one grid away from the aggregation center is set to 0.5, the weight of the grid unit two grids away is set to 0.25, etc. In the weighted average calculation, the characteristic value of each grid unit is multiplied by its corresponding weight, and then all the weighted characteristic values ​​are added and divided by the sum of all weights to obtain the aggregated characteristic value. Starting from the outermost layer of the neighborhood aggregation circle, such weighted average calculation is performed in sequence to the aggregation center to gradually integrate the characteristic information of each grid unit. After completing the calculation of the entire neighborhood aggregation circle, what is finally obtained is the grid aggregation feature of the grid unit. This feature integrates the information of all grid units in the neighborhood and is reasonably weighted according to their distance relationship with the aggregation center. It can more accurately reflect the characteristic state of the grid unit in the entire power grid service spatial structure. At the same time, it also captures the spatial dependence and structural characteristics in the power grid service data, and reasonably integrates the neighborhood grid unit information, so that the grid aggregation feature can more accurately reflect the local and overall relationship of the power grid, further improving the accuracy of the power grid service package configuration.

[0074] Embodiment 2, based on the same inventive concept as the grid service package configuration method combined with grid analysis in the above embodiment, Figure 2 As shown, the present application provides a grid service package configuration system combined with grid analysis, and the system and method embodiments in the present application embodiments are based on the same inventive concept. The system includes:

[0075] The service configuration dimension acquisition module 10 is used to parse the content of the power grid service package to obtain the service configuration dimension.

[0076] The grid granularity determination module 20 determines the grid granularity of the configuration dimension based on the service parameter impact characteristics of the service configuration dimension.

[0077] The segmentation grid determination module 30 is used to perform grid segmentation on each service configuration dimension according to the grid granularity of the configuration dimension to determine the segmentation grid.

[0078] The power feature extraction module 40 is used to analyze each grid unit according to the segmented grid and extract power features, wherein the power features include one or more of power demand, fault characteristics, power facility status, and user demand.

[0079] The grid aggregation feature acquisition module 50 performs feature aggregation on the power features based on the grid granularity of the configuration dimension to obtain grid aggregation features.

[0080] The power grid service package configuration module 60 is used to perform feature projection and configure the power grid service package according to the mapping relationship between the service configuration dimension and the grid aggregation feature.

[0081] Furthermore, the service configuration dimension acquisition module 10 also includes:

[0082] The portrait feature acquisition unit is used to collect the user's electricity usage logs in the target area and obtain the portrait features of the user's electricity usage.

[0083] A user service feedback demand acquisition unit is used to acquire power grid service fault records and user service feedback demands.

[0084] A power grid service demand extraction unit is used to extract user power grid service demands based on the power grid service fault records and user service feedback demands.

[0085] A service configuration dimension acquisition unit, wherein the service configuration dimension acquisition unit is used to take a preset type of a power grid service package as a target, perform a correlation impact analysis based on the profile characteristics of the user's electricity consumption and the user's power grid service needs, and obtain the service configuration dimension based on the correlation impact with the preset type of the power grid service package.

[0086] Furthermore, the service configuration dimension acquisition unit also includes:

[0087] A matching side demand correlation feature acquisition unit is used to pre-process the user's electricity consumption portrait features and the user's power grid service requirements, input them into an association mining model, and obtain matching side demand correlation features.

[0088] A target-side service-related feature acquisition unit is used to pre-process the preset type of the power grid service package, input it into the association mining model, and obtain the target-side service-related feature.

[0089] The association influence acquisition unit is used to input the matching side demand association feature and the target side service association feature into the association mining model to obtain the association influence.

[0090] Furthermore, the matching side demand-related feature acquisition unit further includes:

[0091] An association mining model architecture construction unit, the association mining model architecture construction unit is used to construct an association mining model architecture, including an encoding module, an information association module, and a response generation module, wherein the encoding module is used to identify and divide input information into explicit feature information and implicit feature information, the information association module includes a multi-order iterative association layer to perform association mining on the input feature information, and the response generation module performs correlation calculation and outputs an association result based on the feature description of the association mining; wherein the multi-order iterative association layer includes at least a first-order association layer and a second-order association layer, the first-order association layer extracts the feature according to the explicit feature information, performs feature association mining at the record position, and obtains the first-order mining feature, the second-order association layer obtains the time series feature according to the record position of the feature, and performs feature association mining in combination with the first-order mining feature and the implicit feature information.

[0092] Furthermore, the grid granularity determination module 20 further includes:

[0093] A service parameter impact feature acquisition unit is used to perform configuration impact parameter analysis on the service configuration dimension to obtain service parameter impact features.

[0094] A clustering result acquisition unit is used to perform cluster analysis on the service parameter impact characteristics through a clustering algorithm to obtain a clustering result.

[0095] A grid granularity acquisition unit is used to use the cluster center of the clustering result as the grid unit center, determine the size of the grid unit according to the clustering density of the clustering result, and obtain the grid granularity of the configuration dimension.

[0096] Furthermore, the grid aggregation feature acquisition module 50 also includes:

[0097] The adjacent granularity range setting unit is used to set the adjacent granularity range according to the grid granularity of the configuration dimension.

[0098] A power grid service data graph structure construction unit is used to use grid units as nodes, set connection edges according to the association relationship between grid units, and construct a power grid service data graph structure.

[0099] A neighborhood aggregation circle acquisition unit is used to perform neighborhood extraction on the power grid service data graph structure according to the adjacency granularity range to obtain a neighborhood aggregation circle.

[0100] A grid aggregation feature acquisition unit is used to perform node feature aggregation according to the neighborhood aggregation circle and the connection edge relationship to obtain the grid aggregation feature.

[0101] Furthermore, the grid aggregation feature acquisition unit further includes:

[0102] The aggregation path setting unit is used to use each grid unit as an aggregation center, and according to the aggregation center in the neighborhood aggregation circle, set the aggregation direction and aggregation path according to the distance relationship of the connecting edges.

[0103] The feature weighted averaging unit is used to perform weighted averaging of the features of the grid cells in sequence from far to near from the aggregation center according to the aggregation direction and the aggregation path, so as to obtain the grid aggregation features of the grid cells.

[0104] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0105] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

[0106] This specification and the drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. A method for configuring a power grid service package in combination with grid analysis, characterized in that: The method comprises: Parse the content of the power grid service package to obtain the service configuration dimensions; Determining a grid granularity of a configuration dimension based on a service parameter impact characteristic of the service configuration dimension; Performing grid segmentation on each service configuration dimension according to the grid granularity of the configuration dimension to determine a segmentation grid; According to the segmented grid, each grid unit is analyzed to extract power characteristics, wherein the power characteristics include one or more of power demand, fault characteristics, power facility status, and user demand; Based on the grid granularity of the configuration dimension, performing feature aggregation on the power feature to obtain a grid aggregation feature; According to the mapping relationship between the service configuration dimension and the grid aggregation feature, feature projection is performed to configure the grid service package; Wherein, based on the service parameter impact characteristics of the service configuration dimension, determining the grid granularity of the configuration dimension includes: Performing configuration impact parameter analysis on the service configuration dimension to obtain service parameter impact characteristics; Performing cluster analysis on the service parameter influencing features by using a clustering algorithm to obtain a clustering result; Taking the cluster center of the clustering result as the grid unit center, determining the size of the grid unit according to the clustering density of the clustering result, and obtaining the grid granularity of the configuration dimension; Based on the grid granularity of the configuration dimension, the power feature is aggregated to obtain a grid aggregate feature, including: According to the grid granularity of the configuration dimension, setting the adjacent granularity range; Taking grid units as nodes, setting connection edges according to the association relationship between grid units, and constructing the power grid service data graph structure; Performing neighborhood extraction on the power grid service data graph structure according to the adjacency granularity range to obtain a neighborhood aggregation circle; According to the neighborhood aggregation circle, node features are aggregated according to the connection edge relationship to obtain the grid aggregation features.

2. The method for configuring a power grid service package in combination with grid analysis according to claim 1, characterized in that: Parse the content of the power grid service package to obtain the service configuration dimensions, including: Collect electricity consumption logs of users in the target area to obtain the user's electricity consumption profile features; Obtain grid service fault records and user service feedback requirements; Extracting user grid service demands according to the grid service fault records and user service feedback demands; Taking the preset type of the grid service package as the target, a correlation impact analysis is performed based on the profile characteristics of the user's electricity consumption and the user's grid service needs, and the service configuration dimension is obtained based on the correlation impact with the preset type of the grid service package.

3. The method for configuring a power grid service package in combination with grid analysis according to claim 2, characterized in that: Taking the preset type of grid service package as the target, the correlation impact analysis is performed according to the user's power consumption profile characteristics and the user's grid service needs, including: Preprocessing the user's electricity consumption image features and the user's grid service requirements, inputting them into the association mining model, and obtaining matching side demand association features; Preprocessing the preset type of the power grid service package, inputting it into the association mining model, and obtaining the target side service association feature; The matching-side demand-related features and the target-side service-related features are input into the association mining model to obtain association influence.

4. The method for configuring a power grid service package in combination with grid analysis according to claim 3, characterized in that: Before the input association mining model, it also includes: Constructing an association mining model architecture, including an encoding module, an information association module, and a response generation module, wherein the encoding module is used to identify and divide input information into explicit feature information and implicit feature information, the information association module includes a multi-order iterative association layer to perform association mining on the input feature information, and the response generation module performs correlation calculation based on the feature description of the association mining and outputs the association result; Among them, the multi-order iterative association layer includes at least a first-order association layer and a second-order association layer. The first-order association layer extracts the recorded position of the feature according to the explicit feature information to perform feature association mining to obtain the first-order mining feature. The second-order association layer obtains the time series feature according to the recorded position of the feature, and performs feature association mining in combination with the first-order mining feature and implicit feature information.

5. The method for configuring a power grid service package in combination with grid analysis according to claim 1, characterized in that: According to the neighborhood aggregation circle, node feature aggregation is performed according to the connection edge relationship to obtain the grid aggregation feature, including: Taking each grid unit as the aggregation center, according to the aggregation center in the neighborhood aggregation circle, the aggregation direction and aggregation path are set according to the distance relationship of the connecting edges; According to the aggregation direction and the aggregation path, weighted average of the features of the grid units is performed in sequence from far to near from the aggregation center to obtain the grid aggregation features of the grid units.

6. A grid service package configuration system combined with grid analysis, characterized in that: The system is used to implement the grid service package configuration method combined with grid analysis according to any one of claims 1 to 5, and the system includes: A service configuration dimension acquisition module, wherein the service configuration dimension acquisition module is used to parse the content of the power grid service package to obtain the service configuration dimension; A grid granularity determination module, the grid granularity determination module determines the grid granularity of the configuration dimension based on the service parameter impact characteristics of the service configuration dimension; A segmentation grid determination module, the segmentation grid determination module is used to perform grid segmentation on each service configuration dimension according to the grid granularity of the configuration dimension, and determine the segmentation grid; A power feature extraction module, the power feature extraction module is used to analyze each grid unit according to the segmented grid and extract power features, the power features including one or more of power demand, fault characteristics, power facility status, and user demand; A grid aggregation feature acquisition module, wherein the grid aggregation feature acquisition module performs feature aggregation on the power features based on the grid granularity of the configuration dimension to obtain grid aggregation features; A power grid service package configuration module, the power grid service package configuration module is used to perform feature projection and configure a power grid service package according to a mapping relationship between a service configuration dimension and the grid aggregation feature; Wherein, the grid granularity determination module includes: A service parameter impact feature acquisition unit, the service parameter impact feature acquisition unit is used to perform configuration impact parameter analysis on the service configuration dimension to obtain service parameter impact features; A clustering result acquisition unit, the clustering result acquisition unit is used to perform cluster analysis on the service parameter influence characteristics through a clustering algorithm to obtain a clustering result; A grid granularity acquisition unit, the grid granularity acquisition unit is used to take the cluster center of the clustering result as the grid unit center, determine the size of the grid unit according to the clustering density of the clustering result, and obtain the grid granularity of the configuration dimension; The grid aggregation feature acquisition module includes: An adjacent granularity range setting unit, the adjacent granularity range setting unit is used to set the adjacent granularity range according to the grid granularity of the configuration dimension; A power grid service data graph structure construction unit, wherein the power grid service data graph structure construction unit is used to use grid units as nodes, set connection edges according to association relationships between grid units, and construct a power grid service data graph structure; A neighborhood aggregation circle acquisition unit, the neighborhood aggregation circle acquisition unit is used to extract the neighborhood of the power grid service data graph structure according to the adjacency granularity range to obtain a neighborhood aggregation circle; A grid aggregation feature acquisition unit is used to perform node feature aggregation according to the neighborhood aggregation circle and the connection edge relationship to obtain the grid aggregation feature.

Images