Power grid boundary data processing method and device, power grid boundary data processing system, readable storage medium and program product

Abstract

The application relates to the technical field of power grid data management, and provides a power grid boundary data processing method and device, a power grid boundary data processing system, a readable storage medium and a program product. The method comprises the following steps: mapping each power grid boundary data into standard boundary data; determining the multi-dimensional identifier of each standard boundary data according to the regulation and control object and the dispatch version of the power grid boundary data, so as to obtain compressed boundary data corresponding to the same multi-dimensional identifier; determining the compressed boundary data of the same dispatch version according to the multi-dimensional identifier of the compressed boundary data, so as to obtain a boundary data snapshot of a corresponding dispatch version and store the boundary data snapshot into a snapshot database; when a query request is received, a first boundary data snapshot set is determined in the snapshot database, and then a second boundary data snapshot set is obtained, so as to form a query result. The method supports complex multi-condition screening and multi-granularity time window analysis.

Description

Technical Field

[0001] This application relates to the field of power grid data management technology, and in particular to a method, apparatus, system, computer-readable storage medium, and computer program product for processing power grid boundary data. Background Technology

[0002] With the large-scale integration of new energy sources and the increasing complexity of load characteristics, the changes in the power grid's operating status over time exhibit a significant high-frequency characteristic. Especially in the electricity spot market environment, the grid boundary state is frequently disturbed by factors such as market clearing, forecast correction, and scheduling strategy adjustments, requiring relevant boundary data to be updated on a second-by-second or even millisecond-by-millisecond basis. This places extremely high demands on data storage efficiency, update mechanisms, and historical traceability.

[0003] However, traditional power grid boundary data management methods have many limitations and are difficult to meet the actual needs of whole-grid simulation and retrospective tasks. Specifically, they are also significantly inadequate in terms of multi-dimensional retrieval capabilities, often only allowing linear queries over time. For example, in retrospective analysis, users often want to quickly locate wind power forecast deviation data for a specific region within a certain scheduling cycle or power flow changes for a specific section under a certain clearing cycle, but traditional power grid boundary data management methods cannot meet such precise query needs, severely limiting strategy tracing and simulation efficiency. Summary of the Invention

[0004] Therefore, it is necessary to provide a method, apparatus, system, computer-readable storage medium, and computer program product for processing power grid boundary data, addressing the aforementioned technical problems.

[0005] Firstly, this application provides a method for processing power grid boundary data, including:

[0006] Acquire multi-source heterogeneous power grid boundary data at multiple time points, and map each power grid boundary data to standard boundary data;

[0007] Based on the control object and scheduling version of the power grid boundary data, determine the multidimensional identifier of each standard boundary data;

[0008] Based on the compression algorithm that matches the data characteristics of the standard boundary data, the standard boundary data corresponding to the same multidimensional identifier at multiple time points are compressed to obtain the compressed boundary data corresponding to the same multidimensional identifier.

[0009] Based on the multidimensional identifier of the compressed boundary data, the compressed boundary data of the same scheduling version is identified, so as to obtain the boundary data snapshot of the corresponding scheduling version and store it in the snapshot database; the metadata information of the boundary data snapshot includes the snapshot identifier code, the snapshot generation timestamp and the scheduling version;

[0010] When a query request is received, a first set of boundary data snapshots is determined in the snapshot database based on the correspondence between the region identifier code and the snapshot identifier code, the region identifier code information carried in the query request, and the snapshot identifier code included in the metadata information of the boundary data snapshots.

[0011] Based on the time information and scheduling version information carried in the query request, as well as the metadata information of the boundary data snapshot, including the snapshot generation timestamp and scheduling version, a second boundary data snapshot set is determined from the first boundary data snapshot set;

[0012] Based on the correspondence between the regional identifier code and the regulated object, the target standard boundary data is determined in the second boundary data snapshot set to form the query results.

[0013] Secondly, this application also provides a power grid boundary data processing apparatus, comprising:

[0014] The standard boundary data acquisition module is used to acquire multi-source heterogeneous power grid boundary data at multiple times and map each power grid boundary data to standard boundary data.

[0015] The multidimensional identifier determination module is used to determine the multidimensional identifier of each standard boundary data according to the control object and scheduling version of the power grid boundary data;

[0016] The compressed boundary data acquisition module is used to compress the standard boundary data corresponding to the same multidimensional identifier at multiple time points according to the compression algorithm that matches the data characteristics of the standard boundary data, so as to obtain the compressed boundary data corresponding to the same multidimensional identifier.

[0017] The boundary data snapshot acquisition module is used to determine the compressed boundary data of the same scheduling version based on the multidimensional identifier of the compressed boundary data, so as to obtain the boundary data snapshot of the corresponding scheduling version and store it in the snapshot database; the metadata information of the boundary data snapshot includes the snapshot identifier code, the snapshot generation timestamp and the scheduling version;

[0018] The retrieval module, upon receiving a query request, determines a first set of boundary data snapshots in the snapshot database based on the correspondence between region identifiers and snapshot identifiers, the region identifier information carried in the query request, and the snapshot identifier information included in the metadata information of the boundary data snapshots; determines a second set of boundary data snapshots in the first set of boundary data snapshots based on the time information and scheduling version information carried in the query request, and the snapshot generation timestamp and scheduling version included in the metadata information of the boundary data snapshots; and determines the target standard boundary data in the second set of boundary data snapshots based on the correspondence between region identifiers and control objects to form a query result.

[0019] Thirdly, this application also provides a method for processing power grid boundary data. The processing of the power grid boundary data includes a memory and a processor, the memory storing a computer program, and the processor executing the above-described method.

[0020] Fourthly, this application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, the computer program being executed by a processor using the methods described above.

[0021] Fifthly, this application also provides a computer program product. The computer program product includes a computer program that is executed by a processor using the methods described above.

[0022] The aforementioned methods, apparatus, systems, computer-readable storage media, and computer program products for processing power grid boundary data map each power grid boundary data point to standard boundary data. Based on the control object and scheduling version of the power grid boundary data, a multidimensional identifier for each standard boundary data point is determined, achieving standardized management of multi-source heterogeneous power grid boundary data. This significantly improves the consistency and fusion of power grid boundary data, facilitating cross-system coordination and joint analysis. Using a compression algorithm matching the data characteristics of the standard boundary data, standard boundary data corresponding to the same multidimensional identifier at multiple time points are compressed to obtain compressed boundary data corresponding to the same multidimensional identifier. This significantly reduces storage capacity and network bandwidth overhead while ensuring critical accuracy. Based on the multidimensional identifier of the compressed boundary data, compressed boundary data of the same scheduling version is determined to obtain a snapshot of the boundary data for the corresponding scheduling version and store it in a snapshot database. It possesses a clear scheduling version control system, supporting rapid switching and simulation reconstruction of key historical states. When a query request is received, it determines a first set of boundary data snapshots in the snapshot database based on the correspondence between the area identifier code and the snapshot identifier code, the area identifier code information carried in the query request, and the snapshot identifier code included in the metadata information of the boundary data snapshot. Based on the time information and scheduling version information carried in the query request, and the snapshot generation timestamp and scheduling version included in the metadata information of the boundary data snapshot, it determines a second set of boundary data snapshots from the first set of boundary data snapshots. Based on the correspondence between the area identifier code and the control object, it determines the target standard boundary data in the second set of boundary data snapshots to form the query results. This significantly improves retrieval speed and query flexibility, supports complex multi-condition filtering and multi-granularity time window analysis, and meets the diverse data access needs in power grid scheduling and market operation. Attached Figure Description

[0023] Figure 1 This is a flowchart illustrating a method for processing power grid boundary data in one embodiment;

[0024] Figure 2 This is a schematic diagram of the architecture of a power grid boundary data processing system in one embodiment;

[0025] Figure 3 This is a flowchart illustrating an adaptive compression strategy for boundary data in one embodiment;

[0026] Figure 4 This is a flowchart illustrating the sliding window scrolling update mechanism in one embodiment;

[0027] Figure 5 This is a schematic diagram of a hierarchical storage architecture for processing power grid boundary data in one embodiment;

[0028] Figure 6 This is a flowchart illustrating the event-driven mechanism in one embodiment;

[0029] Figure 7 This is a schematic diagram of a multidimensional index structure and a fast historical data query process in one embodiment;

[0030] Figure 8 This is a structural block diagram of a device for processing power grid boundary data in one embodiment. Detailed Implementation

[0031] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. The terms "comprising" and "having," and any variations thereof, as used in this application, are intended to cover non-exclusive inclusion. The term "multiple" as used in this application refers to two or more. The term "and / or" as used in this application refers to one of the solutions, or any combination of multiple solutions. Mapping to standard boundary data query request query results.

[0032] The method for processing power grid boundary data in this application includes: Figure 1 The steps shown can be applied to power grid boundary data processing systems.

[0033] Step S101: Obtain multi-source heterogeneous power grid boundary data at multiple time points, and map each power grid boundary data to standard boundary data.

[0034] Multi-time, multi-source, heterogeneous power grid boundary data can be a common data source in dispatching operations, including unit output plans in the Operating Control System (OCS), real-time measurements in the Energy Management System (EMS), and short-term load and renewable energy forecasts in the forecasting system. Core fields such as node output, cross-sectional power flow, load level, and renewable energy output forecasts can be extracted and abstracted from the multi-time, multi-source, heterogeneous power grid boundary data, and uniformly mapped into identifiable and parsable standard boundary data (which can be called standard boundary variables) in the multi-dispatch operation model.

[0035] Step S102: Determine the multidimensional identifier of each standard boundary data according to the control object and scheduling version of the power grid boundary data.

[0036] Based on the control objects and scheduling versions of the power grid boundary data, multidimensional identifiers of each standard boundary data can be determined, thereby enabling precise identification and version control for each type of standard boundary data.

[0037] Specifically, a four-dimensional identification system of "control object - timestamp - data type - version number" can be established for each standard boundary data, thereby achieving precise identification and version control for each type of standard boundary data. Among them, the control object can be a power generation unit, substation, or line section; the timestamp can be accurate to the scheduling period or data refresh granularity; the data type can indicate whether the data is a planned value, measured value, or predicted value; and the version number is used to distinguish multiple iterative versions of data at the same point in time due to strategy adjustments and algorithm corrections.

[0038] Step S103: Based on the compression algorithm that matches the data characteristics of the standard boundary data, compress the standard boundary data corresponding to the same multidimensional identifier at multiple time points to obtain the compressed boundary data corresponding to the same multidimensional identifier.

[0039] To improve the storage efficiency of standard boundary data in the whole power grid model while ensuring the reconstruction accuracy of key time-series information, an adaptive data compression algorithm framework can be used to dynamically select a matching compression algorithm based on the data characteristics of the standard boundary data, such as the time-varying characteristics and business sensitivity of the standard boundary data, thereby achieving the best balance between data integrity and real-time performance within limited storage space.

[0040] Step S104: Based on the multidimensional identifier of the compressed boundary data, determine the compressed boundary data of the same scheduling version, so as to obtain the boundary data snapshot of the corresponding scheduling version and store it in the snapshot database; the metadata information of the boundary data snapshot includes the snapshot identifier code, the snapshot generation timestamp and the scheduling version.

[0041] Based on the multidimensional identifier of the compressed boundary data, the compressed boundary data of the same scheduling version can be determined. Based on the compressed boundary data of the same scheduling version, a snapshot of the boundary data of the corresponding scheduling version can be obtained and stored in the snapshot database.

[0042] Specifically, a full boundary data snapshot is generated according to a preset large cycle; within each preset large cycle, multiple preset small cycles are set; in each preset small cycle, incremental boundary data snapshots are generated by extracting the difference boundary data between adjacent boundary data snapshots, thereby saving storage space by only saving the difference boundary data.

[0043] Step S105: When a query request is received, the first set of boundary data snapshots is determined in the snapshot database based on the correspondence between the area identifier code and the snapshot identifier code, the area identifier code information carried in the query request, and the snapshot identifier code included in the metadata information of the boundary data snapshot.

[0044] The query request can specify query conditions, including four dimensions of information such as time range, region, control object, and scheduling version number.

[0045] When a query request is received, based on the area identifier information carried in the query request and the correspondence between the area identifier and the snapshot identifier, the snapshot identifier associated with the area identifier information carried in the query request is determined from the snapshot identifiers included in the metadata information of each boundary data snapshot in the snapshot database. The boundary data snapshot corresponding to the aforementioned associated snapshot identifier is taken as the first boundary data snapshot, thereby determining the first boundary data snapshot set.

[0046] Step S106: Based on the time information and scheduling version information carried in the query request, as well as the metadata information of the boundary data snapshot, including the snapshot generation timestamp and scheduling version, determine the second boundary data snapshot set in the first boundary data snapshot set.

[0047] Based on the time information and scheduling version information carried in the query request, as well as the metadata information of the boundary data snapshot, including the snapshot generation timestamp and scheduling version, a second boundary data snapshot that matches the time information and scheduling version information carried in the query request can be determined from the first boundary data snapshot set, thereby determining the second boundary data snapshot set.

[0048] Step S107: Based on the correspondence between the regional identifier code and the regulated object, determine the target standard boundary data in the second boundary data snapshot set to form the query result.

[0049] Based on the correspondence between the regional identifier code and the regulated object, the target standard boundary data of the regulated object that matches the regional identifier code can be determined in the second boundary data snapshot set, thereby forming the query results.

[0050] In the above-mentioned method for processing power grid boundary data, when a query request is received, a first boundary data snapshot set is determined in the snapshot database based on the correspondence between the area identifier code and the snapshot identifier code, the area identifier code information carried in the query request, and the snapshot identifier code included in the metadata information of the boundary data snapshot. A second boundary data snapshot set is determined based on the time information and scheduling version information carried in the query request, and the snapshot generation timestamp and scheduling version included in the metadata information of the boundary data snapshot. Target standard boundary data is determined in the second boundary data snapshot set based on the correspondence between the area identifier code and the control object to form the query result. This significantly improves retrieval speed and query flexibility, supports complex multi-condition filtering and multi-granularity time window analysis, and meets the diverse data access needs in power grid dispatching and market operation.

[0051] In one embodiment, standard boundary data corresponding to the same multidimensional identifier at multiple time points is compressed using a compression algorithm that matches the data characteristics of the standard boundary data to obtain compressed boundary data corresponding to the same multidimensional identifier. The specific steps are as follows: When the data characteristics of the standard boundary data are that the data redundancy is higher than the redundancy threshold and the error tolerance is lower than the tolerance threshold, the standard boundary data corresponding to the same multidimensional identifier at multiple time points is compressed using a lossless compression algorithm to obtain compressed boundary data corresponding to the same multidimensional identifier; when the data characteristics of the standard boundary data are that the data volatility is higher than the volatility threshold and the error tolerance is higher than the tolerance threshold, the standard boundary data corresponding to the same multidimensional identifier at multiple time points is compressed using a lossy compression algorithm to obtain compressed boundary data corresponding to the same multidimensional identifier; when the data characteristics of the standard boundary data are that the data volatility is higher than the volatility threshold, the uncertainty characteristic is higher than the uncertainty characteristic threshold, and the error tolerance is higher than the tolerance threshold, the standard boundary data corresponding to the same multidimensional identifier at multiple time points is compressed using a lossy quantization compression algorithm to obtain compressed boundary data corresponding to the same multidimensional identifier.

[0052] When the data characteristics of the standard boundary data are that the data redundancy is higher than the redundancy threshold and the error tolerance is lower than the tolerance threshold, the standard boundary data corresponding to the same multidimensional identifier at multiple time points are compressed according to the lossless compression algorithm to obtain the compressed boundary data corresponding to the same multidimensional identifier.

[0053] When the data characteristics of the standard boundary data are that the data volatility is higher than the volatility threshold and the error tolerance is higher than the tolerance threshold, the standard boundary data corresponding to the same multidimensional identifier at multiple time points are compressed according to the lossy compression algorithm to obtain the compressed boundary data corresponding to the same multidimensional identifier.

[0054] When the data characteristics of the standard boundary data are that the data volatility is higher than the volatility threshold, the uncertainty characteristics are higher than the uncertainty characteristic threshold, and the error tolerance is higher than the tolerance threshold, the standard boundary data corresponding to the same multidimensional identifier at multiple time points are compressed according to the lossy quantization compression algorithm to obtain the compressed boundary data corresponding to the same multidimensional identifier.

[0055] In this embodiment, a compression algorithm matching the data characteristics of the standard boundary data is determined based on the data characteristics of the standard boundary data. This algorithm compresses the standard boundary data corresponding to the same multidimensional identifier at multiple time points, resulting in compressed boundary data corresponding to the same multidimensional identifier. This approach can significantly reduce storage capacity and network bandwidth overhead while ensuring critical accuracy.

[0056] In one embodiment, a lossy quantization compression algorithm is used to compress standard boundary data corresponding to the same multidimensional identifier at multiple time points to obtain compressed boundary data corresponding to the same multidimensional identifier. The specific steps are as follows: A multi-level error feedback vector is obtained based on the error distribution at different scales; the feedback errors of each level in the multi-level error feedback vector are weighted and superimposed to obtain a multi-level error feedback superposition vector; the standard boundary data corresponding to the same multidimensional identifier at multiple time points is corrected based on the multi-level error feedback superposition vector to obtain corrected standard boundary data; the quantization step size is dynamically adjusted based on the noise level of the standard boundary data corresponding to the same multidimensional identifier at each time point; the corrected standard boundary data is quantized based on the quantization step size to obtain a standard boundary quantization sequence; and the compressed boundary data corresponding to the same multidimensional identifier is obtained based on the standard boundary quantization sequence.

[0057] It can capture error distributions at different scales and obtain a multi-level error feedback vector as shown in equation (1).

[0058] (1)

[0059] in, This represents the multi-level error feedback vector at time t. Let M represent the error feedback vector at time t, where M represents the error feedback level. Each level is responsible for capturing error distributions at different scales and improving error correction capabilities.

[0060] A time-scale weighted error control mechanism is adopted, which assigns different weights at multiple time levels such as minute and hour, and superimposes the feedback errors of each level in the multi-level error feedback vector according to the weights to obtain a multi-level error feedback superposition vector.

[0061] Based on the multi-level error feedback superposition vector, the standard boundary data corresponding to the same multi-dimensional identifier at multiple time points are corrected to obtain the standard boundary correction data, as shown in Equation (2).

[0062] (2)

[0063] in, This represents the standard boundary correction data at time t. This represents the standard boundary data at time t. This represents the error feedback vector at time t-1, where M represents the error feedback level and the weighting coefficients are... satisfy It can be adaptively adjusted, using the exponential decay weighting method shown in equation (3).

[0064] (3)

[0065] in, This represents the attenuation coefficient.

[0066] Based on the noise level of the standard boundary data at each time step corresponding to the same multidimensional identifier, the quantization step size is dynamically adjusted as shown in Equation (4).

[0067] (4)

[0068] in, This indicates the adjusted quantization step size. Indicates the basic quantization step size. Indicates the adjustment parameter. , This represents the standard boundary data at time t. Indicates the standard deviation of historical data. This represents the signal-to-noise ratio at time t, used to dynamically adjust the quantization accuracy based on the noise level. It represents a small positive number that should not be divided by zero.

[0069] The signal-to-noise ratio is calculated based on the estimation of the signal power and noise power of the standard boundary data within the sliding window, as shown in Equation (5).

[0070] (5)

[0071] in, This represents the signal power at time t. This represents the noise power at time t.

[0072] The standard boundary correction data can be quantized according to the quantization step size to obtain the standard boundary quantization sequence, as shown in equation (6).

[0073] (6)

[0074] in, This represents the standard boundary quantization sequence at time t. This indicates standard boundary correction data. This represents the quantization step size at time t, and round(⋅) represents the rounding operation. Standard boundary quantization sequence. It is an integer encoding, which facilitates subsequent difference and entropy encoding processing. The corresponding reconstructed value is shown in Equation (7).

[0075] (7)

[0076] Compressed boundary data corresponding to the same multidimensional identifier can be obtained based on the standard boundary quantization sequence.

[0077] Specifically, the standard boundary quantization sequence By applying the first-order difference, the incremental data sequence is obtained, as shown in Equation (8).

[0078] (8)

[0079] in, This represents the incremental data sequence at time t (also known as the difference sequence). This represents the standard boundary quantization sequence at time t. This represents the standard boundary quantization sequence at time t-1.

[0080] The incremental data sequence can be converted into a binary compressed bitstream using the Huffman coding algorithm, as shown in equation (9).

[0081] (9)

[0082] in, This represents a binary compressed bitstream.

[0083] Among them, for incremental data sequences Entropy encoding compression is performed using Huffman coding.

[0084] In this embodiment, a multi-level error feedback mechanism effectively mitigates the gradual accumulation of quantization errors in standard boundary data compression and the local shifts caused by sudden anomalies, ensuring numerical fidelity during long-term data compression. Furthermore, a dynamic quantization step size adjustment strategy based on signal-to-noise ratio (SNR) is introduced, which adaptively adjusts the coding precision according to the noise level of the standard boundary data.

[0085] In one embodiment, a multi-level error feedback vector is obtained based on the error distribution at different scales. The specific steps are as follows: When performing the first quantization, a set feedback error is obtained as the multi-level error feedback vector; when performing non-first quantization, the first-level feedback error is obtained based on the difference between the target time series data to be compressed and the corresponding reconstructed data during the last quantization; for each level other than the first level, the level's feedback error is obtained based on the error change of the previous level; and the multi-level error feedback vector is obtained based on the feedback errors of each level.

[0086] When performing the first quantization, the feedback error is set to zero, as shown in equation (10).

[0087] (10)

[0088] The feedback error will be set as a multi-level error feedback vector.

[0089] When performing non-first quantization, the first level of feedback error is obtained based on the difference between the target time series data to be compressed and the corresponding reconstructed data during the last quantization, as shown in Equation (11).

[0090] (11)

[0091] For each level other than the first level, the feedback error of the higher level captures the change of the error of the previous level. Based on the error change of the previous level, the feedback error of the level is obtained, as shown in Equation (12).

[0092] (12)

[0093] Based on the feedback errors at each level, a multi-level error feedback vector is obtained.

[0094] In this embodiment, a multi-level error feedback vector is obtained based on the feedback error at each level, which can effectively track and correct the accumulation of errors at different time scales, thereby significantly improving the accuracy of data quantization.

[0095] In one embodiment, the method provided by this application further includes: capturing boundary data change events through an event-driven mechanism to obtain target power grid boundary data; identifying changes in the target power grid boundary data based on valid boundary data snapshots to obtain a set of changed boundary data; and obtaining an updated boundary data snapshot based on the set of changed boundary data.

[0096] To ensure the timeliness and accuracy of boundary data snapshots in the whole power grid model, an event-driven mechanism can be used to efficiently capture changes in power grid boundary data from multiple heterogeneous external data sources. This means capturing boundary data change events, quickly completing data acquisition, verification, processing, and dynamic updates, and ensuring agile response and stable operation to high-frequency real-time data.

[0097] In the data acquisition phase, an event-driven architecture, combined with modern data middleware technologies such as Kafka message queues, can be used to achieve real-time proactive monitoring and data push of multiple heterogeneous data sources, such as scheduling and control systems, energy management systems, and new energy forecasting systems.

[0098] After capturing boundary data change events, the corresponding power grid boundary data can be obtained as the target power grid boundary data. The intelligent change identification module can identify changes in the target power grid boundary data based on valid boundary data snapshots, resulting in a set of changed boundary data. This identified set of changed boundary data is dynamically written to a rolling cache structure based on a sliding time window. The caching mechanism automatically overwrites expired boundary data snapshots and eliminates invalid information, maintaining the up-to-dateness and compactness of the boundary data snapshots.

[0099] In this embodiment, boundary data change events are captured through an event-driven mechanism to obtain target power grid boundary data, thereby obtaining an updated boundary data snapshot. This enables boundary data snapshot updates at the second level or even shorter time scales, meeting the stringent requirements of real-time review and dispatch optimization in the electricity spot market.

[0100] In one embodiment, changes to the target power grid boundary data are identified based on valid boundary data snapshots to obtain a set of changed boundary data. The specific steps are as follows: a valid boundary data snapshot is obtained based on the full snapshot of the boundary data closest to the time when the boundary data change event occurred, and incremental snapshots of the boundary data between the time when the full snapshot of the boundary data was generated and the time when the boundary data change event occurred; a field-level comparison is performed between the valid boundary data snapshot and the target power grid boundary data to obtain a set of changed boundary data.

[0101] A valid boundary data snapshot is obtained by taking the full snapshot of the boundary data closest to the time when the boundary data change event occurred, and the incremental snapshots of the boundary data between the time when the full snapshot of the boundary data was generated and the time when the boundary data change event occurred.

[0102] Field-level comparison is performed between valid boundary data snapshots and target power grid boundary data to achieve precise comparison. By classifying events by type (such as plan modification, measurement data update, and forecast adjustment) and verifying by timestamp and version number, the actual change range of boundary data is accurately determined, redundant information is filtered out, and subsequent processing is ensured to be executed only for the actual changed boundary data, effectively reducing system load and improving overall response efficiency.

[0103] A valid boundary data snapshot can be defined as... The target power grid boundary data is Then, the field-level change judgment is as shown in equation (13).

[0104] (13)

[0105] in, This indicates a change in boundary data.

[0106] For each field i, let the change threshold be... Then change the boundary data set As shown in equation (14).

[0107] (14)

[0108] Only for The middle field is processed further to reduce invalid updates.

[0109] In this embodiment, a valid boundary data snapshot is obtained based on the full snapshot of the boundary data closest to the time when the boundary data change event occurs, and the incremental snapshot of the boundary data between the time when the full snapshot of the boundary data is generated and the time when the boundary data change event occurs. By comparing the valid boundary data snapshot with the target power grid boundary data at the field level, a more accurate set of changed boundary data can be obtained.

[0110] like Figure 2As shown, this application provides a power grid boundary data processing system, which includes a unified data access layer, a compression processing module, a version snapshot management module, a multi-dimensional indexing and querying module, and an event-driven update module. The unified data access layer supports multi-protocol access and standardized parsing from various power grid dispatch boundary data sources (such as actual output, forecast data, equipment parameters, and planning instructions), serving as the system's data entry point. The compression processing module dynamically selects algorithms such as differential encoding, moving average, and quantization compression based on data characteristics and business scenarios to adaptively compress the accessed data to reduce storage load. Compressed data is synchronously transmitted to the version snapshot management module, which uses a sliding window and incremental storage strategy to generate multi-granularity time snapshots and implements traceable management of historical states through a version control mechanism. The indexing and querying module constructs a multi-dimensional index system based on time, equipment, and region, supporting users to quickly query historical boundary data or trace back to any point in time snapshot as needed. The event-driven update module monitors key dispatch events in real time (such as load forecast updates and equipment switching), automatically triggering the refresh, anomaly detection, and cleaning processes for relevant boundary data. The modules work together through standard interfaces to form a unified closed-loop data management system, ensuring that the system has high compression ratio, low latency and strong consistency.

[0111] like Figure 3 As shown, this application employs an adaptive compression strategy for standard boundary data compression. This strategy is an intelligent decision-making mechanism designed to flexibly select the optimal compression algorithm to balance storage efficiency and data accuracy, taking into account the diversity and dynamic changes of standard boundary data. First, the input standard boundary data is categorized into two main types: measured data and predicted data. For measured data, its rate of change within the sliding time window is calculated to assess the data's stability and volatility. When the rate of change is low, lossless compression algorithms, such as differential coding or Zstandard compression, are prioritized to ensure data integrity and high-precision storage. If the rate of change is high, the decision is based on the stringency of data accuracy requirements in the current business context. For scenarios with strong consistency requirements, lossless compression continues; for scenarios where some accuracy loss is permissible, lossy compression strategies, such as lossy coding or adaptive quantization coding, are selected to effectively reduce storage space. For predicted data, the system intelligently decides whether to use high-precision lossless compression or lossy methods such as segmented compression or quantization compression, based on its error tolerance and actual business needs, to adapt to the inherent uncertainty of new energy prediction data. The entire process supports dynamic adjustment of compression strategies and feeds the final decision results back to the compression processing module, achieving closed-loop optimization and continuous improvement, effectively enhancing the overall performance and application adaptability of power grid boundary data management.

[0112] like Figure 4As shown, the sliding window scrolling update mechanism details the key steps and inter-module collaboration processes in the system's reception, processing, and storage of power grid boundary data, where User represents the user. First, external data sources continuously push boundary data change events to the event-driven module, which serves as the entry point for the entire update process, responsible for capturing and transmitting the latest data change signals in real time. Upon receiving the data, the event-driven module transmits the target power grid boundary data to the data verification module for multi-level data integrity and validity verification, including format verification, range detection, and timestamp consistency confirmation. During verification, the anomaly handling module intervenes, automatically correcting (e.g., interpolation, filtering) or marking any detected anomalies as abnormal. If the anomaly is severe, a manual review process is triggered to ensure data quality and system stability.

[0113] Once the target power grid boundary data passes verification, the event-driven module notifies the version control module. The version control module is responsible for maintaining the version number system and incremental dependencies, ensuring that each boundary data snapshot update forms a unique and ordered version chain. Subsequently, the updated boundary data snapshot, after version confirmation, is written to a sliding window cache structure. The sliding window, with a fixed length as its boundary, automatically overwrites the oldest boundary data snapshot and evicts expired data, ensuring that the cache always retains the latest and continuous time-series data. While ensuring high-speed read and write, the cached data is synchronized to the underlying persistent storage system in real time, achieving secure data storage and long-term retention. The system supports multi-threaded concurrent writes, and the sliding window and version control module work closely together to achieve synchronous updates and version consistency guarantees in high-frequency data stream environments. Simultaneously, the query module quickly indexes the required boundary data snapshots through the version control module and requests historical boundary data snapshots from the storage system, supporting second-level historical version queries and runtime replays. The overall mechanism ensures an end-to-end closed loop from data generation, verification, version management, cache maintenance to storage query, efficiently handling the high-frequency updates and multi-version management needs of power grid boundary data, achieving real-time performance, stability, and reliability of the system.

[0114] A hierarchical storage architecture for processing power grid boundary data, such as Figure 5 As shown, this diagram depicts the organization and flow of boundary data at different storage levels within the system, meeting the diverse needs of power grid dispatching for data timeliness and historical traceability. The system first receives and caches power grid boundary data change events from multiple heterogeneous data sources at the real-time layer. The real-time layer employs a high-speed in-memory database or in-memory caching technology, supporting millisecond-level write and query responses to ensure that the latest boundary data can be quickly captured and processed instantly. Data is continuously updated at the real-time layer, and a fixed-length time series is maintained through a sliding window mechanism to ensure the temporal integrity and continuity of the data. As time progresses, some data in the real-time layer gradually sinks to the short-term layer according to preset time window rules.

[0115] The short-term tier primarily stores grid boundary data from the last few days to weeks. It utilizes a high-performance time-series database or fast-access solid-state drives (SSDs) for storage, ensuring both rapid data retrieval and a degree of storage optimization. The short-term tier not only preserves complete version snapshots and incremental differences but also supports multi-dimensional index structures, facilitating scheduling review and short-term analysis. Data in the short-term tier is received from the real-time tier through periodic batch or incremental writes, and based on data access frequency and historical value, a portion of the data is migrated to the long-term tier.

[0116] The long-term tier handles the archiving of massive amounts of historical boundary data, typically employing large-capacity, low-cost hard disk drives or cloud storage solutions. The long-term tier archives data in monthly, quarterly, or even annual units, with compressed and optimized storage formats to reduce space consumption, and efficient indexed metadata supports cross-period historical queries. Access to long-term tier data has relatively high latency, primarily serving non-real-time scenarios such as in-depth analysis, scheduling review, and market research. The overall tiered storage architecture, through the organic collaboration of the real-time, short-term, and long-term tiers, achieves a balance between high-speed response and massive storage, effectively supporting the full lifecycle management and multi-scenario application needs of power grid boundary data.

[0117] Event-driven mechanisms (also known as event-driven dynamic data update and cleansing mechanisms) such as Figure 6The diagram illustrates in detail how the system efficiently updates and ensures the quality of power grid boundary data after receiving key events such as adjustments to the power grid market plan and equipment switching. The process begins with the multi-source heterogeneous data interface capturing boundary data change events via message middleware. These events are then pushed to the system's event processing module, which categorizes and identifies them, determining the scope and priority of data to be updated based on the event type. Subsequently, the system pulls or receives the latest boundary data from the corresponding data source, entering the intelligent change identification phase. This phase accurately compares the old and new data, filters redundant updates, and ensures that subsequent processing is only performed on genuine changes, reducing system load. Next, the updated data is written to a sliding window-based rolling cache. The caching mechanism automatically overwrites expired information, ensuring data timeliness and continuity. During this process, an anomaly detection module is activated in parallel, using statistical thresholds, business rules, and machine learning models to monitor data quality and identify anomalous data such as mutations, missing values, and logical conflicts. For minor anomalies, the system automatically triggers a data cleaning process, including interpolation correction and smoothing filtering, to ensure data continuity and accuracy. For severe anomalies, alarms are generated and the system initiates a manual review process. After data cleaning is completed, the updated boundary data snapshot is synchronously pushed to the version snapshot management and multi-dimensional indexing modules, enabling version control and efficient retrieval of boundary data snapshots, supporting subsequent scheduling simulation and market analysis. This mechanism achieves closed-loop automated management from event triggering to data cleaning and updating, significantly improving the real-time performance, accuracy, and intelligent level of power grid boundary data and system operation.

[0118] Multidimensional index structure and fast historical data query process, as follows Figure 7 As shown, this diagram illustrates a data indexing system built upon key dimensions such as time, region, and equipment, along with its corresponding retrieval paths. The structure uses time as the main thread, encompassing multiple time granularities including scheduling cycles, clearing cycles, and operation cycles, forming a multi-layered time index that supports rapid data location within any time range. The region-level index categorizes boundary data by region based on the geographical division of the power grid or administrative regions, facilitating cross-regional data comparison, analysis, and filtering. The equipment-level index classifies data according to the control objects within the power grid, such as generator units, AC / DC lines, and new energy forecasting units, refining the data to the specific equipment level to meet the needs of thematic analysis. Through the combined use of these three-dimensional indexes, query requests can accurately match the corresponding time period, geographical region, and equipment, significantly improving data retrieval efficiency and accuracy, meeting the requirements for second-level response and historical review. The overall structure adopts a tag-based storage model, combined with time-series database technology, ensuring tight coupling between the index and actual data, enabling efficient and flexible multi-condition filtering and aggregation queries, supporting diverse application scenarios such as scheduling simulation, market analysis, and anomaly tracking.

[0119] To better understand the above method, the following describes in detail an application embodiment of the power grid boundary data processing method of this application, which can be applied to a power grid boundary data processing system.

[0120] With the rapid development of the electricity spot market, the system dispatching model is undergoing a profound transformation. It is gradually shifting from a plan-driven approach to a data-driven approach based on dynamic control and strategy optimization. Under this trend, grid dispatching operations heavily rely on accurate historical operational status replays, high-frequency simulation analysis based on actual boundary conditions, and intelligent model calculations supporting predictive decision-making. The prerequisite for achieving these functions is the construction of a unified boundary data management system covering the entire grid, capable of dynamic updates and retrospection. Grid boundary data refers to the dataset used to define the input and output boundary conditions of the control model, encompassing key control parameters such as the unit output plans and dispatching instructions of the central / provincial dispatch centers, real-time power flow at inter-regional AC / DC sections, power forecasts and actual output of renewable energy, market clearing results, bidding curves, and reserve plans. This data is not only widely sourced and heterogeneous in format, but also updated frequently and changes state rapidly, serving as the fundamental dataset supporting operational simulations and market behavior replays.

[0121] With the large-scale integration of new energy sources and increasingly complex load characteristics, the changes in the power grid's operating status over time exhibit a significant high-frequency characteristic. Especially in the spot market environment, the grid boundary state is frequently disturbed by factors such as market clearing, forecast correction, and scheduling strategy adjustments, requiring relevant boundary data to be updated on a second-by-second or even millisecond-by-millisecond basis. This places extremely high demands on data storage efficiency, update mechanisms, and historical traceability.

[0122] However, traditional power grid boundary data processing systems have many limitations and are no longer sufficient to meet the actual needs of full power grid simulation and review tasks. Firstly, regarding storage mechanisms, traditional systems mostly employ a fixed-period full-data storage strategy, lacking sliding window or incremental compression mechanisms. This results in high redundancy and large storage overhead for high-frequency data (such as renewable energy output and real-time power flow), and low efficiency in reading long-term historical data, severely restricting the ability to quickly build models and reconstruct scenarios. Secondly, regarding dynamic update mechanisms, traditional systems generally rely on periodic data collection, failing to respond in real-time to sudden events such as plan modifications and power flow mutations. This leads to data version lag and inconsistent status, consequently affecting the accuracy of simulations of operational scenarios built based on this data.

[0123] Furthermore, there are significant shortcomings in multi-dimensional retrieval capabilities. Traditional systems often only allow linear queries by time, making it difficult to support multi-dimensional combined indexes based on region, object, time, and type. For example, in retrospective analysis, users often want to quickly locate "wind power forecast deviation data for a certain region within a certain scheduling cycle" or "power flow changes at a certain cross section under a certain clearing cycle," but traditional systems struggle to meet such precise query needs, severely limiting the efficiency of strategy tracing and simulation.

[0124] More critically, there is currently a lack of a systematic version management mechanism. Scheduling decisions often involve multiple version stages (such as original manual planning, market clearing adjustments, and real-time scheduling corrections). The boundary data corresponding to each stage differs. If version information cannot be fully recorded and managed, it becomes difficult to accurately reconstruct the actual boundary conditions during post-mortem analysis, thus affecting the judgment of the rationality of scheduling behavior and the evaluation of optimization strategies. Furthermore, in terms of data quality assurance, traditional systems lack the ability to proactively identify and correct abnormal data. Issues such as excessive deviations in new energy forecasts and sudden changes in real-time power flow are often mistakenly taken as real inputs, causing distortion in simulation results.

[0125] Therefore, based on traditional technologies, there is an urgent need to propose a method for processing power grid boundary data applicable to full power grid simulation scenarios, possessing the following capabilities: First, supporting unified access and abstract modeling of multi-source heterogeneous boundary data; second, possessing efficient rolling storage and incremental snapshot functions to improve storage efficiency; third, possessing a clear data version control system to support accurate backtracking of operational status at any point in time; fourth, establishing a multi-dimensional efficient indexing mechanism to achieve data retrieval with second-level response; and fifth, integrating data compression and anomaly repair modules to ensure the timeliness, accuracy, and integrity of boundary data. The establishment of such a method will become a core foundational capability supporting intelligent scheduling, operational simulation, and strategy optimization in a spot market environment.

[0126] The specific steps for processing power grid boundary data are as follows:

[0127] Step 1: Unified modeling of power grid boundary data:

[0128] The unified boundary data structure is based on common data sources in scheduling operations—including unit output plans in the operation control planning system, real-time measurements in the energy management system, and short-term load and renewable energy forecasts in the forecasting system. It extracts and abstracts core fields such as node output, cross-sectional power flow, load level, and renewable energy output forecasts, mapping them uniformly to identifiable and parsable standard boundary data in the model. To effectively support data backtracking and multi-version management required for scheduling scenario reconstruction and result review, a four-dimensional identification system of "control object - timestamp - data type - version number" is further introduced to achieve precise identification and version control for each type of standard boundary data. The control object can be a power generation unit, substation, or line cross-section; the timestamp can be accurate to the scheduling period or data refresh granularity; the data type can indicate whether the data is a planned value, measured value, or forecast value; and the version number is used to distinguish multiple iterative versions of data at the same point in time due to strategy adjustments and algorithm corrections. This four-dimensional identification system ensures that the system can quickly and accurately locate the required boundary data when performing tasks such as operational review and real-time control scenario switching, and provides a consistent data interface for different algorithm modules, thereby improving the reconfigurability and data traceability of the overall control process.

[0129] Step 2, Rolling Storage Mechanism and Snapshot Building:

[0130] In this step, boundary data snapshots are periodically built according to a set time granularity (e.g., every 1 minute or 5 minutes), and a sliding window structure is used for rolling storage to achieve the overwriting of new snapshot data and the automatic elimination of old data. Each snapshot record is embedded with complete metadata information, including timestamp, data source system identifier, scheduling version number, etc., to ensure that historical data is traceable and reproducible, providing basic support for subsequent simulation calculations or fault backtracking.

[0131] To meet the real-time access requirements of large-scale boundary data, such as high-frequency updates, accurate tracking, and low-latency access, in power grid control processes, a sliding window-based rolling storage mechanism and snapshot construction method are proposed. This method combines dynamic data writing, multi-threaded concurrency control, and version management to achieve efficient data storage and rapid response. The mechanism aims to periodically collect and encapsulate the state of the entire network boundary data at set time granularities (e.g., 1 minute or 5 minutes), constructing multi-version, multi-type boundary data snapshots, and ensuring stable and controllable system resource consumption even as the data volume continues to grow.

[0132] The specific steps of this mechanism are as follows:

[0133] Specifically, a sequence of boundary data snapshots on the time axis can be defined. for:

[0134] (15)

[0135] in, Indicates a point in time A complete snapshot of boundary data is generated (e.g., every 1 minute or 5 minutes), containing the boundary data status of all controlled objects at that moment.

[0136] Each boundary data snapshot Defined as an ordered quadruple: , This represents the boundary data matrix, which records the various types of boundary values ​​of each controlled object at that moment. Indicates the data source identifier (such as operation control planning system, energy management system, and forecasting system); This represents the scheduling version number, used to identify the business scenario corresponding to the boundary data snapshot; It represents a collection of metadata information, including collection timestamps, data precision levels, data tags, etc.

[0137] Set the cache time window length W (e.g., 24 hours, 7 days) according to actual needs, and initialize the sliding window data structure. Pre-allocate cache space and create a timestamp-based index to support fast location and updates. Configure a cache capacity limit to avoid memory overflow and set an eviction policy threshold. Capture boundary data change events through an event-driven mechanism. The system performs validity verification and change identification on newly received data items, filtering out the set of boundary data that have actually changed. Cache write operations are performed only on this collection to reduce invalid data writes.

[0138] To avoid the accumulation of storage burden over time, the system uses a sliding window structure to maintain the data snapshot queue. Let the sliding window length be W, meaning the system only retains the most recent W snapshots, and any exceeding this limit are automatically discarded. This process can be expressed as follows:

[0139] This process can be expressed as follows:

[0140] (16)

[0141] in, Represents an active boundary data snapshot. Indicates time A snapshot of the boundary data.

[0142] When new boundary data snapshot After it is generated, the window slides forward one position:

[0143] (17)

[0144] The above mechanism can ensure that the latest W versions of boundary data snapshots are always maintained, which not only meets the data integrity requirements of tasks such as real-time data analysis, control simulation and review, but also ensures that storage resources are under control.

[0145] Selected valid change data Write to the cache according to timestamp t. If the write time exceeds the start time of the current window. Then the sliding window will start sliding forward, updating the window range to...

[0146] (18)

[0147] At the same time, old data outside the window is automatically discarded. :

[0148] (19)

[0149] This method cleans up expired data, frees up cache space, and ensures that the cache always keeps the latest data.

[0150] Furthermore, to support efficient retrieval and multi-source comparison, the boundary data entries in each snapshot are structured using a unified four-dimensional index (controlling object ID, timestamp, data type, and version number), forming the following index function:

[0151] (20)

[0152] Where 'o' represents the control object, 'type' represents the data type (such as planned value, measured value, predicted value), and 'v' represents the scheduling version number. This indicates the specific boundary data value retrieved under this index.

[0153] In this way, when new version data is written, if the version number is updated, the old version is automatically replaced, while historical version copies are retained for backtracking queries and multi-version comparisons, allowing the system to quickly restore any historical state during retrospective analysis. This snapshot construction and rolling storage mechanism realizes temporal compressed storage of boundary data, version traceability, and consistent access, and is an important foundational module for subsequent model invocation, algorithm training, and simulation reproduction.

[0154] Multi-threaded concurrent write control is adopted, and cache writes are processed by multiple threads to cope with high-frequency data inflow. A read-write lock mechanism is designed to ensure thread safety: read operations are executed concurrently to ensure fast access response; write operations are locked to ensure mutual exclusion of write operations at the same time; an optimistic locking mechanism is adopted to reduce lock waiting and improve concurrency efficiency.

[0155] Furthermore, in order to improve the storage efficiency of boundary data in the whole power grid model while ensuring the reconstruction accuracy of key time-series information, this system introduces an adaptive data compression algorithm framework. It dynamically selects an appropriate compression strategy based on the time-varying characteristics and business sensitivity of the boundary data, thereby achieving the best balance between data integrity and real-time performance within limited storage space.

[0156] This method dynamically matches compression strategies to different types of boundary data, enabling intelligent selection and adjustment of compression modes. Based on the volatility, trend, and error tolerance characteristics of the data itself, it flexibly employs various compression methods, including time moving average, difference coding, and segmented compression, to achieve an optimal balance between data accuracy and storage cost.

[0157] Specifically, for data such as scheduling instructions, control boundaries, and actual power output, although the data redundancy is high, the accuracy requirements are still extremely high in applications such as simulation backtesting, consistency verification of control boundaries, and fine-grained scheduling optimization. Furthermore, these data are typically used for metering backtesting and fine-grained scheduling control. Therefore, lossless compression algorithms can be prioritized to ensure that accuracy is not lost while improving the compression ratio and avoiding the accidental deletion or misjudgment of details during critical periods. For wind / solar power forecast data and high-frequency real-time power flow measurements, lossy compression algorithms are used. Real-time power flow measurement data fluctuates wildly, has a high sampling frequency, and contains a large amount of short-term disturbances and redundant information. Using lossy compression algorithms can remove invalid high-frequency fluctuations within the allowable accuracy error range, significantly improving compression efficiency and storage performance, while still retaining the main trend characteristics. This meets the engineering accuracy requirements of simulation analysis and scheduling review, and significantly reduces the storage and transmission burden while ensuring the preservation of key fluctuation characteristics. For new energy forecast data, which fluctuates wildly and has significant uncertainty, its main use is in trend identification and power assessment, allowing for some information loss within the control error range. To address the high uncertainty of new energy forecast data, a lossy quantization compression algorithm based on a dynamic error feedback mechanism can be selected. This algorithm utilizes non-uniform quantization and adaptive step size adjustment to achieve efficient compression while maintaining data compression ratio and accuracy.

[0158] The adaptive data compression algorithm framework adopts a segmented compression strategy as follows:

[0159] (twenty one)

[0160] in, This represents the standard boundary data to be compressed. This represents a segment of standard boundary data to be compressed. express Standard boundary data for each time step, for each segment of standard boundary data Based on its changing trend, a compression strategy (lossy compression, lossless compression, or no compression) is selected, and segment header information and compression type metadata are recorded to achieve flexible decoding.

[0161] In addition, metadata is added to each compressed data record, including start and end timestamps (Time Range), data source system, data version number (Version ID), and compression strategy type (such as Delta compression, Quantized compression, and Raw compression).

[0162] The adaptive data compression algorithm framework can be formally represented as:

[0163] (twenty two)

[0164] in, This indicates the data to be compressed. This indicates the selected compression algorithm. Let E represent the set of candidate compression algorithms, E represent the compression error, and R represent the storage ratio. This represents the compression precision weighting coefficient, which can automatically and dynamically adjust and select the optimal compression strategy based on data characteristics. .

[0165] To address the high uncertainty of new energy forecast data, an improved adaptive quantization compression algorithm is designed, balancing data compression efficiency with accuracy preservation. This algorithm is based on a dynamic error feedback mechanism, utilizing non-uniform quantization and adaptive step size adjustment to achieve efficient compression.

[0166] Because new energy forecast data exhibits strong volatility and has a relatively high tolerance for errors, a lossy quantization compression strategy can be introduced to reduce the burden of data storage and transmission. To improve the compression effect, an improved method is proposed: First, a non-uniform quantization strategy is combined to better adapt to the distribution characteristics of the forecast data, prioritizing the retention of accuracy in the small fluctuation range while appropriately relaxing accuracy requirements for the large fluctuation range; second, an error feedback mechanism is introduced to cumulatively correct errors generated during the quantization process, thereby effectively suppressing error propagation and distortion amplification; finally, by dynamically adjusting the quantization step size Δt, the accuracy and compression ratio during the compression process can be adaptively optimized based on the data characteristics. Specific steps are shown in the above embodiment.

[0167] This algorithm effectively mitigates the gradual accumulation of quantization errors during time series compression and local shifts caused by sudden anomalies through a multi-level error feedback mechanism, ensuring numerical fidelity during long-term data compression. Building upon this, a dynamic quantization step size adjustment strategy based on signal-to-noise ratio (SNR) is introduced, enabling the system to adaptively adjust coding precision according to the noise level of real-time data. This is particularly suitable for the high uncertainty and high-frequency fluctuation characteristics commonly found in new energy forecasting data. Simultaneously, a time-scale weighted error control mechanism is employed, assigning different weights at multiple time levels, such as minute and hourly, to differentiate between short-term and long-term errors, thereby improving overall stability and predictive usability.

[0168] In terms of compression methods, the system combines differential coding and entropy coding to significantly improve the data compression rate. Differential processing can remove trend components, and entropy coding further approximates the theoretically optimal coding length, meeting the storage resource constraints of new energy data with a second-level update frequency. The overall algorithm structure is clear, with low computational complexity, making it easy to deploy and implement in existing boundary data acquisition and control systems. It is a new energy boundary data management solution that balances real-time performance, compression rate, and decoding accuracy.

[0169] Step 3, Snapshot Version Management and Differentiated Storage:

[0170] This step introduces a version-oriented boundary data storage strategy. Differences in boundary data between consecutive snapshots are extracted and stored, recording only the changes compared to the previous version snapshot, significantly saving storage space. Simultaneously, complete snapshots are retained for key scheduling nodes (such as day-ahead clearing and real-time scheduling) to support scenario reconstruction. By integrating a version control mechanism, the entire change trajectory of each clearing result and plan adjustment is fully recorded, ensuring that boundary data snapshot versions are controllable and modifications are traceable.

[0171] To ensure that the multi-scheduling business model in the power grid dispatching system can achieve complete backtracking and stable recovery of power grid boundary data at critical moments, a key dispatching node snapshot mechanism is established to generate a full boundary data snapshot during the core operation phase. This mechanism is designed for dispatching business scenarios with high-frequency changes and high reliability requirements, and is particularly targeted at the following key nodes:

[0172] Important scheduling cycle start and end points: such as switching from day-ahead planning to real-time scheduling, or starting hourly rolling scheduling; significant scheduling instruction change times: such as changes in operating mode, start-up and shutdown of large units, and adjustments to inter-network connection schemes; market clearing result release times: such as day-ahead clearing, real-time clearing, or adjustment plan release nodes.

[0173] At the aforementioned key nodes, a complete snapshot of the boundary data will be generated, denoted as... It is defined as being in the scheduling node A full copy of the time boundary data set, i.e.:

[0174] (twenty three)

[0175] Each boundary data snapshot not only contains all boundary data items, but also includes metadata information, including: snapshot generation timestamp. Snapshot version number Source system identifier Snapshot generation trigger type (Periodic / Event-driven).

[0176] Overall boundary data snapshot structure This can be formally represented as:

[0177] (twenty four)

[0178] To balance data integrity and storage efficiency, the snapshot mechanism combines a hybrid strategy of periodic triggering and event-driven triggering:

[0179] Specific event triggering function Formalizable definition:

[0180] (25)

[0181] in, Indicates the magnitude of changes in scheduling instructions. This indicates that a threshold value is set. This indicates whether the market clearing power has changed.

[0182] Once a boundary data snapshot is generated, it is immediately written to a highly reliable storage structure and linked to the version control chain in a chained manner, serving as the anchor point for restoring all incremental versions after that moment. If a rollback to any version is needed subsequently... Efficient restoration can be achieved through the following process:

[0183] (26)

[0184] in, express A snapshot of boundary data at a given time. express Incremental boundary data snapshot at time step.

[0185] To meet the high requirements of the power grid dispatching system for "traceability, rollback, and auditability" of power grid boundary data, this application establishes a version control mechanism to systematically manage all critical data changes. This mechanism revolves around four core functions: "version identification, change tracking, dependency maintenance, and historical reproducibility," ensuring high consistency and controllability in data management even with frequent updates to power grid boundary data and highly dynamic scenarios.

[0186] Unified Version Identifier Management: Each update to power grid boundary data triggers the automatic generation of a version number, forming a unique version identifier. The version number adopts a multi-level encoding format (e.g., year-scheduling cycle-change sequence number), facilitating version organization and rapid location across time and scenarios. This identifier is used throughout the entire data processing process, ensuring version consistency across all processing stages.

[0187] Change log mechanism: Detailed operation logs are generated for each update of power grid boundary data. Records include: timestamp, operator identification (ID) or system source, modified data item identifier, old and new value summaries, and modification reason or scheduling trigger event tag. Change logs are stored in a chained structure, strongly correlated with the corresponding version number, forming a complete data evolution chain, providing a reliable foundation for anomaly tracing and scheduling analysis.

[0188] Version dependency and incremental recovery mechanism: Maintain the dependency graph between boundary data snapshot versions and incremental data. Each target version can be represented as a boundary data snapshot plus several subsequent increments. This structure ensures that when it is necessary to restore the historical running state, the entire boundary data can be reconstructed efficiently and accurately through "snapshot + incremental playback".

[0189] Version switching and data restoration interface: To facilitate use by dispatchers, market analysts and system self-diagnosis modules, a standardized version switching interface is provided, which supports one-click switching to any historical version, automatic loading of corresponding snapshots and incremental sequences, and selection to restore the data state to a specific point in time, thereby improving the flexibility and accuracy of scheduling scenario review.

[0190] In summary, this version control mechanism, through structured modeling and process-oriented management, improves the timeliness, completeness, and traceability of data management, significantly enhancing the robustness and operational transparency of the new energy dispatching system under changing operating environments. The differentiated storage strategy combining incremental storage and snapshots offers significant advantages over traditional full-data storage methods. This strategy not only drastically reduces storage capacity requirements, especially in scenarios with high-frequency updates of boundary data, but also significantly improves data write and read efficiency, supporting second-level real-time updates and historical backtracking operations. Furthermore, this solution supports flexible historical version management, meeting diverse needs for dispatching review and optimization analysis, while also being compatible with multi-source heterogeneous data structures, facilitating the integration of boundary data from different systems, and greatly enhancing the system's applicability and scalability.

[0191] Step 4: Multidimensional Indexing and Efficient Retrieval Mechanism

[0192] A multi-dimensional composite index system was constructed, comprising four core dimensions: time, region, regulated object, and version number. The time dimension covers different scheduling cycles, clearing cycles, and operating cycles, supporting multi-granularity time range queries. The region dimension is based on geographical and administrative divisions, enabling cross-regional data location. The regulated object dimension is subdivided into multiple categories such as generating units, lines, and new energy units, facilitating thematic analysis of specific objects. The version number dimension ensures version accuracy during data review. The flexible combination of these multi-dimensional indexes meets users' needs for accurate data location and rapid response in complex query scenarios.

[0193] For underlying data management, an advanced time-series database is adopted as the storage engine, fully leveraging its optimized design for time-series data. The time-series database features high-throughput batch write capabilities, fast time-based retrieval, automatic compression and multi-level aggregation functions, and rich tag query support. Time, region, and control object information in the multi-dimensional index are mapped to tags and fields in the time-series database, enabling efficient multi-dimensional filtering and aggregation queries, significantly improving overall read / write performance and system scalability.

[0194] Furthermore, considering the multi-timescale characteristics of power grid dispatching and market operation data, the system supports a multi-period indexing mechanism, including dispatching cycle (minute-level), clearing cycle (market node), and operation cycle (hourly, daily-level) indexes, allowing flexible switching between different time granularities. Combined with parallel query execution, incremental index dynamic updates, and caching mechanisms, it achieves sub-second response times for complex joint queries. This retrieval framework is widely applicable to various scenarios such as real-time operation review, market strategy optimization, and new energy fluctuation analysis, significantly improving the intelligence and precision of power grid boundary data management.

[0195] Step 5: Real-time dynamic update mechanism:

[0196] This step introduces an event-driven data refresh mechanism to achieve data linkage and updates with external systems. The system can proactively listen for boundary data update events (such as plan modifications, measurement refreshes, forecast adjustments, etc.) and trigger a rolling cache update process after recognizing data changes. Simultaneously, it performs anomaly detection on the incoming data. Once abrupt changes or missing items are identified, it automatically triggers data cleaning and quality verification processes to ensure that the boundary data input to the system is complete, accurate, and real-time.

[0197] To ensure the timeliness and accuracy of boundary data snapshots in the whole power grid model, an event-driven mechanism can be used to efficiently capture changes in power grid boundary data from multiple heterogeneous external data sources. This means capturing boundary data change events, quickly completing data acquisition, verification, processing, and dynamic updates, and ensuring agile response and stable operation to high-frequency real-time data.

[0198] In the data acquisition phase, an event-driven architecture, combined with modern data middleware technologies such as Kafka message queues, can be used to achieve real-time proactive monitoring and data push of multiple heterogeneous data sources, such as scheduling and control systems, energy management systems, and new energy forecasting systems.

[0199] After capturing boundary data change events, the corresponding power grid boundary data can be obtained as the target power grid boundary data. The intelligent change identification module can identify changes in the target power grid boundary data based on valid boundary data snapshots, resulting in a set of changed boundary data. This identified set of changed boundary data is dynamically written to a rolling cache structure based on a sliding time window. The caching mechanism automatically overwrites expired boundary data snapshots and eliminates invalid information, maintaining the up-to-dateness and compactness of the boundary data snapshots.

[0200] To ensure high-quality boundary data snapshots, an automatic anomaly detection and cleaning module was implemented. Utilizing statistical thresholds, machine learning anomaly detection models, and power system business rules, it automatically identifies anomalies such as data mutations, missing data, and logical conflicts (e.g., power flow exceeding limits, negative output). For minor anomalies, the module automatically corrects them using interpolation and smoothing filtering algorithms, significantly reducing the impact of anomalies on subsequent model simulations and scheduling decisions. For severely anomalous data, it marks the data and triggers a manual review process to ensure the reliability and accuracy of the overall data chain. Simultaneously, the module continuously monitors data integrity and time series continuity, ensuring seamless connection and overall consistency of the boundary data stream.

[0201] Statistical anomaly detection is based on sliding window data distribution, defining the mean of field i within a time window W. with standard deviation for:

[0202] (27)

[0203] in, This represents the target power grid boundary data.

[0204] Anomaly detection criteria:

[0205] (28)

[0206] in, This represents the abnormal threshold coefficient.

[0207] Automatic cleaning uses interpolation or smoothing functions. Correcting outliers:

[0208] (29)

[0209] in, This represents the corrected target power grid boundary data.

[0210] In summary, the event-driven mechanism possesses significant advantages such as low latency, high reliability, intelligence, and efficient resource utilization. It can achieve boundary data updates on a second-level or even shorter timescale, meeting the stringent requirements of real-time review and dispatch optimization in the electricity spot market. This mechanism is compatible with multi-source heterogeneous data and flexibly adapts to various data interfaces and formats, effectively ensuring the timeliness and quality of boundary data across the entire power grid. It serves as a key technical support and core guarantee for the digital transformation of the power system, intelligent dispatch, and accurate market operation review.

[0211] The above-mentioned method for processing power grid boundary data has the following innovative features:

[0212] (1) Unified modeling mechanism for boundary data of multi-source heterogeneous systems: The four-element identifier system of "regulation object-timestamp-data type-version number" is introduced to realize the semantic fusion and structural unification of multi-source high-frequency boundary data in the scheduling system. (2) Rolling snapshot mechanism with sliding window control: Data snapshots are generated periodically using the sliding window structure, and the snapshot length is controlled by time granularity, which effectively balances data real-time performance and storage resource overhead. (3) Differentiated storage + snapshot version chain mechanism: To address the content redundancy problem between snapshot data, incremental differential storage and key point complete snapshot retention strategy are introduced to improve snapshot storage efficiency and data traceability. (4) Fast historical data retrieval mechanism based on multi-dimensional index structure: A high-performance index structure is built according to the dimensions of time, region, regulation object, etc., and combined with time series database optimization, to realize second-level query and accurate backtracking of historical boundary data. (5) Event-driven data update and anomaly verification mechanism: The event stream of heterogeneous systems is accessed by using Kafka / Webhook and other methods, and anomaly detection and cleaning are performed at the same time as data update to ensure the accuracy and reliability of boundary data.

[0213] The aforementioned method for processing power grid boundary data achieves standardized management of multi-source heterogeneous data by constructing a unified boundary data model and a four-dimensional identification system, greatly improving data consistency and fusion, and facilitating cross-system coordination and joint analysis. Secondly, the use of a rolling storage mechanism and sliding window snapshot construction effectively controls storage scale, avoids data redundancy, and ensures the system always maintains the latest boundary data set, meeting the need for second-level data updates. Simultaneously, the combination of snapshot version management and differentiated storage strategies enables efficient data storage and access, supports rapid historical version backtracking and scenario restoration, and improves the accuracy and efficiency of scheduling simulation and operational review. Furthermore, the integration of multi-dimensional indexing and time-series database technology significantly improves data retrieval speed and query flexibility, supports complex multi-condition filtering and multi-granularity time window analysis, and meets the diverse data access needs in power grid scheduling and market operation. Finally, the real-time dynamic update mechanism, based on event-driven design, enables agile response and efficient processing of external boundary data changes. Combined with automatic cleaning and verification of abnormal data, it ensures the accuracy and stability of system data. This invention effectively solves the problems of delay, redundancy and traceability in boundary data management in the current scheduling system, and provides solid technical support for intelligent scheduling, accurate review and spot market analysis.

[0214] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps. The steps in different embodiments can be freely combined as needed, and all non-contradictory solutions formed by such combinations are within the scope of protection of this application.

[0215] In one exemplary embodiment, such as Figure 8 As shown, this application provides a power grid boundary data processing apparatus, wherein:

[0216] The standard boundary data acquisition module 801 is used to acquire multi-source heterogeneous power grid boundary data at multiple times and map each power grid boundary data to standard boundary data.

[0217] The multidimensional identifier determination module 802 is used to determine the multidimensional identifier of each standard boundary data according to the control object and scheduling version of the power grid boundary data;

[0218] The compressed boundary data acquisition module 803 is used to compress the standard boundary data corresponding to the same multidimensional identifier at multiple time points according to the compression algorithm that matches the data characteristics of the standard boundary data, so as to obtain the compressed boundary data corresponding to the same multidimensional identifier.

[0219] The boundary data snapshot acquisition module 804 is used to determine the compressed boundary data of the same scheduling version based on the multidimensional identifier of the compressed boundary data, so as to obtain the boundary data snapshot of the corresponding scheduling version and store it in the snapshot database; the metadata information of the boundary data snapshot includes the snapshot identifier code, the snapshot generation timestamp and the scheduling version;

[0220] The retrieval module 805 is configured to, upon receiving a query request, determine a first set of boundary data snapshots in the snapshot database based on the correspondence between the region identifier code and the snapshot identifier code, the region identifier code information carried in the query request, and the snapshot identifier code included in the metadata information of the boundary data snapshot; determine a second set of boundary data snapshots in the first set of boundary data snapshots based on the time information and scheduling version information carried in the query request, and the snapshot generation timestamp and scheduling version included in the metadata information of the boundary data snapshots; and determine the target standard boundary data in the second set of boundary data snapshots based on the correspondence between the region identifier code and the controlled object, so as to form a query result.

[0221] In one embodiment, the compressed boundary data acquisition module 803 is further configured to: obtain a multi-level error feedback vector based on error distributions at different scales; weight the feedback errors of each level in the multi-level error feedback vector to obtain a multi-level error feedback superposition vector; correct the standard boundary data corresponding to the same multi-dimensional identifier at multiple time points based on the multi-level error feedback superposition vector to obtain standard boundary correction data; dynamically adjust the quantization step size based on the noise level of the standard boundary data corresponding to the same multi-dimensional identifier at each time point; quantize the standard boundary correction data based on the quantization step size to obtain a standard boundary quantization sequence; and obtain the compressed boundary data corresponding to the same multi-dimensional identifier based on the standard boundary quantization sequence.

[0222] In one embodiment, the compressed boundary data acquisition module 803 is further configured to: when the data characteristics of the standard boundary data are that the data redundancy is higher than the redundancy threshold and the error tolerance is lower than the tolerance threshold, compress the standard boundary data corresponding to the same multidimensional identifier at multiple time points according to a lossless compression algorithm to obtain the compressed boundary data corresponding to the same multidimensional identifier; when the data characteristics of the standard boundary data are that the data volatility is higher than the volatility threshold and the error tolerance is higher than the tolerance threshold, compress the standard boundary data corresponding to the same multidimensional identifier at multiple time points according to a lossy compression algorithm to obtain the compressed boundary data corresponding to the same multidimensional identifier; when the data characteristics of the standard boundary data are that the data volatility is higher than the volatility threshold, the uncertainty characteristic is higher than the uncertainty characteristic threshold, and the error tolerance is higher than the tolerance threshold, compress the standard boundary data corresponding to the same multidimensional identifier at multiple time points according to a lossy quantization compression algorithm to obtain the compressed boundary data corresponding to the same multidimensional identifier.

[0223] In one embodiment, the compression boundary data acquisition module 803 is further configured to: acquire a set feedback error as a multi-level error feedback vector when performing the first quantization; when performing non-first quantization, obtain the first-level feedback error based on the difference between the target time series data to be compressed and the corresponding reconstructed data during the last quantization; for each level other than the first level, obtain the feedback error of the level based on the error change of the previous level; and obtain a multi-level error feedback vector based on the feedback errors of each level.

[0224] In one embodiment, the apparatus further includes a boundary data snapshot update module, configured to: capture boundary data change events through an event-driven mechanism to obtain target power grid boundary data; identify changes in the target power grid boundary data based on valid boundary data snapshots to obtain a set of changed boundary data; and obtain an updated boundary data snapshot based on the set of changed boundary data.

[0225] In one embodiment, the boundary data snapshot update module is further configured to: obtain a valid boundary data snapshot based on the full boundary data snapshot closest to the time of the boundary data change event, and incremental boundary data snapshots between the time of generation of the full boundary data snapshot and the time of the boundary data change event; and perform a field-level comparison between the valid boundary data snapshot and the target power grid boundary data to obtain a set of changed boundary data.

[0226] Each module in the aforementioned power grid boundary data processing device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in hardware within or independently of the processor in the power grid boundary data processing system, or stored in software within the memory of the power grid boundary data processing system, so that the processor can call and execute the corresponding operations of each module.

[0227] In one embodiment, a power grid boundary data processing system is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0228] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0229] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0230] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0231] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this application. The above embodiments only illustrate several implementation methods of this application, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make several modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for processing power grid boundary data, characterized in that, The method includes: Acquire multi-source heterogeneous power grid boundary data at multiple time points, and map each power grid boundary data to standard boundary data; Based on the control object and scheduling version of the power grid boundary data, determine the multidimensional identifier of each standard boundary data; When the data characteristics of the standard boundary data are that the data redundancy is higher than the redundancy threshold and the error tolerance is lower than the tolerance threshold, the standard boundary data corresponding to the same multidimensional identifier at multiple time moments are compressed according to the lossless compression algorithm to obtain the compressed boundary data corresponding to the same multidimensional identifier. When the data characteristics of the standard boundary data are that the data volatility is higher than the volatility threshold and the error tolerance is higher than the tolerance threshold, the standard boundary data corresponding to the same multidimensional identifier at multiple times is compressed according to the lossy compression algorithm to obtain the compressed boundary data corresponding to the same multidimensional identifier. When the data characteristics of the standard boundary data are that the data volatility is higher than the volatility threshold, the uncertainty characteristics are higher than the uncertainty characteristic threshold, and the error tolerance is higher than the tolerance threshold, the standard boundary data corresponding to the same multidimensional identifier at multiple times is compressed according to the lossy quantization compression algorithm to obtain the compressed boundary data corresponding to the same multidimensional identifier. Based on the multidimensional identifier of the compressed boundary data, the compressed boundary data of the same scheduling version is identified, so as to obtain the boundary data snapshot of the corresponding scheduling version and store it in the snapshot database; the metadata information of the boundary data snapshot includes the snapshot identifier code, the snapshot generation timestamp and the scheduling version; When a query request is received, a first set of boundary data snapshots is determined in the snapshot database based on the correspondence between the region identifier code and the snapshot identifier code, the region identifier code information carried in the query request, and the snapshot identifier code included in the metadata information of the boundary data snapshots. Based on the time information and scheduling version information carried in the query request, as well as the metadata information of the boundary data snapshot, including the snapshot generation timestamp and scheduling version, a second boundary data snapshot set is determined from the first boundary data snapshot set; Based on the correspondence between the regional identifier code and the regulated object, the target standard boundary data is determined in the second boundary data snapshot set to form the query results.

2. The method according to claim 1, characterized in that, Based on a lossy quantization compression algorithm, standard boundary data at multiple time points corresponding to the same multidimensional identifier are compressed to obtain compressed boundary data corresponding to the same multidimensional identifier, including: Based on the error distribution at different scales, a multi-level error feedback vector is obtained; The feedback errors of each level in the multi-level error feedback vector are superimposed according to their weights to obtain the multi-level error feedback superposition vector. Based on the multi-level error feedback superposition vector, the standard boundary data corresponding to the same multi-dimensional identifier at multiple time moments are corrected to obtain the standard boundary correction data. The quantization step size is dynamically adjusted based on the noise level of the standard boundary data at each time step corresponding to the same multidimensional identifier. The standard boundary correction data is quantized according to the quantization step size to obtain the standard boundary quantization sequence; Based on the standard boundary quantization sequence, compressed boundary data corresponding to the same multidimensional identifier is obtained.

3. The method according to claim 2, characterized in that, The process of obtaining multi-level error feedback vectors based on error distributions at different scales includes: When performing the first quantization, the set feedback error is obtained and used as a multi-level error feedback vector. When performing non-first quantization, the first level of feedback error is obtained based on the difference between the target time series data to be compressed and the corresponding reconstructed data at the time of the last quantization. For each level other than the first level, the feedback error of the level is obtained based on the error change of the previous level. Based on the feedback errors at each level, a multi-level error feedback vector is obtained.

4. The method according to claim 1, characterized in that, The method further includes: The target power grid boundary data is obtained by capturing boundary data change events through an event-driven mechanism; Based on the snapshot of valid boundary data, the target power grid boundary data is modified to obtain a set of modified boundary data. Based on the changed boundary data set, an updated boundary data snapshot is obtained.

5. The method according to claim 4, characterized in that, The step of identifying changes to the target power grid boundary data based on a snapshot of valid boundary data yields a set of changed boundary data, including: A valid boundary data snapshot is obtained by taking the full snapshot of the boundary data closest to the time when the boundary data change event occurred, and the incremental snapshot of the boundary data between the time when the full snapshot of the boundary data was generated and the time when the boundary data change event occurred. A field-level comparison is performed between the valid boundary data snapshot and the target power grid boundary data to obtain a set of changed boundary data.

6. A device for processing power grid boundary data, characterized in that, The device includes: The standard boundary data acquisition module is used to acquire multi-source heterogeneous power grid boundary data at multiple times and map each power grid boundary data to standard boundary data. The multidimensional identifier determination module is used to determine the multidimensional identifier of each standard boundary data according to the control object and scheduling version of the power grid boundary data; The compressed boundary data acquisition module is used to compress standard boundary data corresponding to the same multidimensional identifier at multiple time points using a lossless compression algorithm when the data characteristics of the standard boundary data are that the data redundancy is higher than the redundancy threshold and the error tolerance is lower than the tolerance threshold. When the data characteristics of the standard boundary data are that the data volatility is higher than the volatility threshold and the error tolerance is higher than the tolerance threshold, the module compresses the standard boundary data corresponding to the same multidimensional identifier at multiple time points using a lossy compression algorithm. When the data characteristics of the standard boundary data are that the data volatility is higher than the volatility threshold, the uncertainty characteristic is higher than the uncertainty characteristic threshold, and the error tolerance is higher than the tolerance threshold, the module compresses the standard boundary data corresponding to the same multidimensional identifier at multiple time points using a lossy quantization compression algorithm. The boundary data snapshot acquisition module is used to determine the compressed boundary data of the same scheduling version based on the multidimensional identifier of the compressed boundary data, so as to obtain the boundary data snapshot of the corresponding scheduling version and store it in the snapshot database; the metadata information of the boundary data snapshot includes the snapshot identifier code, the snapshot generation timestamp and the scheduling version; The retrieval module, upon receiving a query request, determines a first set of boundary data snapshots in the snapshot database based on the correspondence between region identifiers and snapshot identifiers, the region identifier information carried in the query request, and the snapshot identifier information included in the metadata information of the boundary data snapshots; determines a second set of boundary data snapshots in the first set of boundary data snapshots based on the time information and scheduling version information carried in the query request, and the snapshot generation timestamp and scheduling version included in the metadata information of the boundary data snapshots; and determines the target standard boundary data in the second set of boundary data snapshots based on the correspondence between region identifiers and control objects to form a query result.

7. The apparatus according to claim 6, characterized in that, The compressed boundary data acquisition module is also used to obtain multi-level error feedback vectors based on error distributions at different scales; to superimpose the feedback errors of each level in the multi-level error feedback vectors according to weights to obtain a multi-level error feedback superposition vector; to correct the standard boundary data at multiple times corresponding to the same multi-dimensional identifier based on the multi-level error feedback superposition vectors to obtain standard boundary correction data; and to dynamically adjust the quantization step size based on the noise level of the standard boundary data at each time step corresponding to the same multi-dimensional identifier. Based on the quantization step size, the standard boundary correction data is quantized to obtain a standard boundary quantization sequence; based on the standard boundary quantization sequence, compressed boundary data corresponding to the same multidimensional identifier is obtained.

8. A system for processing power grid boundary data, comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 5.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 5.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 5.

Images