Search

CN-121705699-B - Hongmon-based power distribution panoramic intelligent management and control decision data mining and processing method

CN121705699BCN 121705699 BCN121705699 BCN 121705699BCN-121705699-B

Abstract

The invention provides a method for mining and processing intelligent control decision data of a power distribution panorama based on hong, which relates to the technical field of intelligent decision and data mining of power systems, and comprises the steps of carrying out feature analysis and state mapping on an optimized feature set to obtain an operation state vector; the method comprises the steps of carrying out multi-objective optimization calculation based on an operation state vector to obtain a primary control parameter set, carrying out instruction coding and strategy fusion on the primary control parameter set to obtain a power distribution control instruction, transmitting the power distribution control instruction to a corresponding terminal, and adopting a time sequence synchronization and hash check mechanism to ensure the sequence consistency and the integrity of instruction transmission. The invention constructs a full-flow closed loop from data safety acquisition, high-efficiency processing and intelligent mining to accurate decision making and reliable execution, and realizes panoramic intelligent control and self-adaptive optimization of the power distribution system under the drive of data.

Inventors

  • GUAN TIANYU

Assignees

  • 北京开鸿数字科技发展有限公司

Dates

Publication Date
20260508
Application Date
20251125

Claims (10)

  1. 1. The intelligent management and control decision data mining and processing method for the power distribution panorama based on the hong Monte is characterized by comprising the following steps of: Step 100, accessing and collecting multi-source heterogeneous original data of a power distribution terminal through a hong-Monte distributed soft bus, carrying out protocol analysis and unified packaging on the multi-source heterogeneous original data to obtain a standardized data packet, and carrying out encryption and equipment identity authentication on the standardized data packet to obtain a safe encrypted data stream; step 200, decrypting the secure encrypted data stream to obtain a decrypted data stream, carrying out multi-source protocol analysis on the decrypted data stream to obtain preliminary structured data, carrying out data cleaning and space-time alignment on the preliminary structured data to obtain a data sequence with consistent space time; Step 300, inputting a standardized data stream into a pre-trained lightweight data mining model, and executing load classification prediction, abnormal state detection and new energy consumption capability analysis at a hong Monte edge end to obtain a basic feature set; generating dynamic compensation coefficients according to the partition configuration strategy, and carrying out real-time calibration on the basic feature set by utilizing the dynamic compensation coefficients to obtain an optimized feature set; Step 400, carrying out feature analysis and state mapping on the optimized feature set to obtain an operation state vector, carrying out multi-objective optimization calculation based on the operation state vector to obtain a preliminary control parameter set, and carrying out instruction coding and strategy fusion on the preliminary control parameter set to obtain a power distribution management and control instruction; and 500, issuing a power distribution control instruction to a corresponding terminal, and adopting a time sequence synchronization and hash check mechanism to ensure the sequence consistency and the integrity of instruction transmission.
  2. 2. The method for mining and processing intelligent management and control decision data for electric power distribution panorama based on hong Meng according to claim 1, wherein step 100 comprises: collecting multisource heterogeneous original data through a hong Mongolian distributed soft bus; Identifying multi-source heterogeneous original data, and distinguishing IEC61850 protocol data, MQTT telemetry data and custom JSON format data contained in the multi-source heterogeneous original data; Respectively carrying out multi-protocol analysis on IEC61850 protocol data, MQTT telemetry data and custom JSON format data, and identifying and extracting effective fields; performing semantic reconstruction and format unification on the effective fields, and packaging the effective fields into a canonical data packet with unified time sequence identification and data source identification; And performing redundant compression and length alignment on the standard data packet to obtain a standardized data packet.
  3. 3. The method for mining and processing intelligent management and control decision data for electric power distribution panorama based on hong Meng in claim 2, wherein step 200 comprises: decrypting and checking the integrity of the safe encrypted data stream to obtain a decrypted data stream; Carrying out multi-source protocol analysis on the decrypted data stream, and identifying a data source and a format to obtain primary structured data; performing time stamp calibration and equipment space topology matching on the effective data set to obtain a data sequence with consistent space and time; The method comprises the steps of carrying out transmission characteristic analysis on a data sequence with consistent space and time, determining a fragmentation configuration parameter, carrying out dynamic fragmentation on the data sequence based on the fragmentation configuration parameter to obtain fragmented data units, adding redundancy check marks to the fragmented data units, and carrying out sequential coding to obtain data units to be recombined; and carrying out recombination and integrity verification on the data unit to be recombined to obtain a standardized data stream.
  4. 4. A method for mining and processing intelligent control decision data for electric power distribution panorama based on hong Meng according to claim 3, wherein step 300 comprises: Inputting the standardized data stream into a lightweight data mining model fused with a pre-trained pruning decision tree and a deep forest to obtain a primary feature map; based on the primary feature mapping, carrying out load mode classification through a hierarchical discrimination structure of a pruning decision tree to obtain a load state feature vector; The load state feature vector and the primary feature map are input to a multi-granularity scanning mechanism of a depth forest together, and abnormal mode identification is carried out to obtain an abnormal index vector; based on the load state feature vector and the abnormal index vector, carrying out new energy output characteristic analysis through a regression analysis unit of a lightweight data mining model to obtain a digestion potential feature vector; And carrying out feature vectorization processing on the load state feature vector, the abnormal index vector and the digestion potential feature vector to obtain vectorized features, and carrying out weighted fusion on the vectorized features to obtain a basic feature set.
  5. 5. The method for intelligent management and control decision data mining and processing for electric power distribution panorama based on hong Meng in claim 4, further comprising the steps of: Defining coordinate axes of feature dimensions based on the basic feature set and establishing a multi-dimensional feature space; mapping the basic feature set to the multidimensional feature space to form feature data point distribution; Performing cluster analysis on the characteristic data point distribution by adopting a spatial clustering method based on density, and identifying a high-density region of data distribution to generate a partition identification set; calculating the centroid position and boundary range of each partition in the partition identification set, and determining the space topological relation between the partitions; and integrating the partition centroid position, the boundary range and the space topological relation to obtain a partition configuration strategy.
  6. 6. The method for intelligent management and control decision data mining and processing for electric power distribution panorama based on hong Meng in claim 5, further comprising the steps of: The feature data subsets corresponding to the partitions are extracted according to the partition configuration strategy, and statistical analysis is carried out on the feature data subsets of each partition, so that the mean value and the variance of the feature values in the partitions are calculated respectively; calculating the ratio of the variance to the mean value as a partition stability index based on the mean value and the variance of the characteristic values in the partition; according to the partition stability index, a dynamic compensation coefficient is obtained through a preset mapping relation; Receiving the dynamic compensation coefficient and organizing the dynamic compensation coefficient into a partition compensation vector according to a partition configuration strategy; Based on the partition configuration strategy, establishing a corresponding relation between each feature vector in the basic feature set and a partition compensation vector; according to the corresponding relation between each feature vector and the partition compensation vector, performing point multiplication operation on each feature vector and the corresponding partition compensation vector, and completing the weighted calibration of the feature values to obtain the feature vector after weighted calibration; and integrating all the weighted and calibrated feature vectors to obtain an optimized feature set.
  7. 7. The method for intelligent management and control decision data mining and processing for electric power distribution panorama based on hong Meng in claim 6, wherein step 400 comprises: performing feature dimension analysis on the optimized feature set, extracting key operation features and establishing a feature weight matrix; performing weighted aggregation on the optimized feature set based on the feature weight matrix to obtain an operation state vector; based on the running state vector, constructing a multi-objective optimization function taking power supply reliability, economic running cost and new energy consumption rate as core targets, and setting weight coefficients of the targets according to running requirements; Carrying out iterative solution on the weighted multi-objective optimization function through a self-adaptive decision algorithm to obtain a candidate control parameter set; Performing constraint condition verification on the candidate control parameter set, and screening feasible control parameters meeting the safe operation boundary of the power grid; based on the weight coefficient of each target, weighting comprehensive evaluation is carried out on the feasible control parameters to obtain a primary control parameter set; Performing instruction code conversion on the preliminary control parameter set, and mapping the preliminary control parameter set into a standardized equipment control instruction; And carrying out strategy fusion on the equipment control instruction based on a preset distribution network topology structure so as to obtain a distribution management and control instruction of coordinated control.
  8. 8. The method for intelligent management and control decision data mining and processing for electric power distribution panorama based on hong Meng in claim 7, wherein step 500 comprises: receiving the power distribution control instruction, and sequencing and packaging the power distribution control instruction according to a preset instruction execution time sequence to obtain an instruction execution queue; The method comprises the steps of adding a time stamp and a serial number to each instruction in an instruction execution queue to form an instruction stream with a time sequence mark, calculating the instruction stream with the time sequence mark based on an SHA-256 hash algorithm to generate a digital fingerprint of instruction data as a first re-checking code, calculating the instruction data by adopting a CRC32 cyclic redundancy check algorithm to generate a second re-checking code, and obtaining an instruction data packet based on the first re-checking code and the second re-checking code; Establishing communication connection with a target terminal through a hong-Monte-involved distributed network, and sequentially sending instruction data packets according to the order of an instruction execution queue; in the process of sending the instruction data packets, acquiring transmission state feedback of each instruction data packet; based on the transmission state feedback, identifying the instruction data packet with failed transmission, starting an automatic retransmission mechanism, identifying the instruction data packet with successful transmission, and performing integrity verification on the instruction data packet with successful transmission by using a check code fed back by a receiving end to obtain a data integrity confirmation signal; Obtaining a confirmation signal of successful transmission according to the data integrity confirmation signal; and based on the confirmation signal of successful transmission, receiving confirmation feedback of completion of terminal execution, updating the instruction execution state, and completing the whole-flow closed-loop management issued by the power distribution control instruction.
  9. 9. A hong-based power distribution panorama intelligent management decision data mining and processing system implementing the method of any one of claims 1 to 8, comprising: The system comprises an acquisition module, a protocol analysis and unified packaging module, a security encryption module, a data processing module and a data processing module, wherein the acquisition module is used for accessing and acquiring multi-source heterogeneous original data of a power distribution terminal through a hong-Mongolian distributed soft bus; The system comprises a standardized module, a data cleaning module, a data processing module and a data processing module, wherein the standardized module is used for decrypting the safe encrypted data stream to obtain a decrypted data stream, carrying out multi-source protocol analysis on the decrypted data stream to obtain preliminary structured data, carrying out data cleaning and space-time alignment on the preliminary structured data to obtain a data sequence with consistent space time; The optimization module is used for inputting the standardized data stream into a pre-trained lightweight data mining model, and executing load classification prediction, abnormal state detection and new energy consumption capability analysis at the hong-Meng edge end to obtain a basic feature set; The system comprises a fusion module, a primary control parameter set, a power distribution management and control instruction, a power distribution management and control module and a power distribution management module, wherein the fusion module is used for carrying out feature analysis and state mapping on an optimized feature set to obtain an operation state vector; And the execution module is used for issuing the power distribution control instruction to the corresponding terminal, and ensuring the sequence consistency and the integrity of instruction transmission by adopting a time sequence synchronization and hash check mechanism.
  10. 10. A computing device, comprising: One or more processors; Storage means for storing one or more programs which when executed by the one or more processors cause the one or more processors to implement the method of any of claims 1 to 8.

Description

Hongmon-based power distribution panoramic intelligent management and control decision data mining and processing method Technical Field The invention relates to the technical field of intelligent decision and data mining of power systems, in particular to a method for mining and processing intelligent control decision data of power distribution panorama based on hong Monte. Background As urban distribution networks continue to expand in size and distributed energy permeability continues to increase, traditional distribution network management and control systems face serious challenges in terms of data processing. The existing system generally adopts a centralized data processing architecture, and has obvious technical bottlenecks in the aspects of multi-source heterogeneous data fusion, edge computing capability adaptation, dynamic data calibration and the like. In the data acquisition layer, IEC61850 protocol data, MQTT telemetry data and custom JSON format data generated by a power distribution terminal lack of unified access specifications, so that serious time sequence dislocation occurs in the protocol analysis process, actual operation data shows that the data packet loss rate is up to 12% when multi-source data are accessed concurrently, the time consumption of a data preprocessing link is more than 65% of the total processing time, and the data processing efficiency is possibly restricted. In the edge computing level, the existing lightweight data mining model is limited by the memory capacity of an edge terminal, which is generally less than or equal to 256MB, and cannot effectively support real-time reasoning of a complex algorithm, when input data exceeding 50 dimensions are processed, model reasoning delay is suddenly increased from 200ms to 1.2s, the memory overflow probability reaches 23%, and a system is forced to upload 70% of original data to cloud processing, so that the abnormal detection response time is possibly beyond a safety critical value of 15 seconds. In the aspect of characteristic engineering, the traditional method has insufficient adaptability to time-varying characteristics of the power distribution running state, when the fluctuation of new energy output exceeds 30% of rated capacity, a characteristic drift phenomenon occurs in a characteristic extraction model trained based on historical data, the accuracy of load classification is reduced by 28%, a dynamic compensation mechanism based on the real-time running state is lacking in a system, the characteristic vector cannot be adaptively calibrated according to the actual working condition of a power grid, and obvious deviation between a control decision and actual demands is possibly caused. Disclosure of Invention The invention aims to solve the technical problem of providing a method for mining and processing intelligent control decision data of a power distribution panoramic view based on hong Mongolian, which constructs a full-flow closed loop from data safety collection, efficient processing, intelligent mining to accurate decision and reliable execution, and realizes panoramic intelligent control and self-adaptive optimization of a power distribution system under data driving. In order to solve the technical problems, the technical scheme of the invention is as follows: in a first aspect, a method for mining and processing intelligent management and control decision data of a power distribution panorama based on hong Monte, the method comprising: The method comprises the steps of accessing and collecting multi-source heterogeneous original data of a power distribution terminal through a hong-Mongolian distributed soft bus, carrying out protocol analysis and unified packaging on the multi-source heterogeneous original data to obtain a standardized data packet, and carrying out encryption and equipment identity authentication on the standardized data packet to obtain a safe encrypted data stream; The method comprises the steps of decrypting a safe encrypted data stream to obtain a decrypted data stream, carrying out multi-source protocol analysis on the decrypted data stream to obtain primary structured data, carrying out data cleaning and space-time alignment on the primary structured data to obtain a space-time consistent data sequence, and carrying out fragmentation and recombination processing on the space-time consistent data sequence to obtain a standardized data stream; The method comprises the steps of inputting a standardized data stream into a pre-trained lightweight data mining model, performing load classification prediction, abnormal state detection and new energy consumption capability analysis on a hong Monte-Meng edge end to obtain a basic feature set, constructing a multi-dimensional feature space based on the basic feature set, and partitioning the multi-dimensional feature space to obtain a partition configuration strategy; Performing feature analysis and state mapping on the optimized feature set to