Grid Hierarchical and Zonal Intelligent Analysis Method Based on Regulation Cloud

Abstract

The present invention discloses a method for intelligent analysis of power grid hierarchical partitioning based on a control cloud, and relates to the technical field of intelligent analysis of power grids. The existing power grid analysis methods have low analysis accuracy and fixed traditional partitioning methods, and are difficult to adapt to complex power grid operating environments. The present invention includes the following steps: under the control cloud architecture, data is collected for power grids at different levels, and the short-term characteristics, long-term trend characteristics, and spatial correlation characteristics of power grid data are extracted using spatiotemporal feature extraction functions, and according to the distribution differences of power grid data characteristics at different levels, a hierarchical adaptive clustering algorithm is used to perform adaptive fusion of data; a hierarchical network structure is used to perform intelligent analysis of power grid hierarchical partitioning; potential power grid anomalies are identified by constructing a state transition matrix; according to power grid data at different levels, the operating status of each level is analyzed, and intelligent optimization decisions and control execution are performed based on abnormal characteristics and global impacts, thereby improving the automation and intelligence level of power grid control.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid intelligent analysis, and in particular to a control cloud-based power grid hierarchical and regional intelligent analysis method. Background Art

[0002] As the scale of power grids expands and the penetration rate of new energy sources increases, traditional power grid control models are unable to meet the needs of intelligence. This invention introduces a control cloud architecture to achieve intelligent hierarchical and regional analysis, thereby improving the efficiency and stability of power grid operation. However, existing technologies still face the following challenges:

[0003] Traditional power grid control models have limited computing power and are unable to efficiently process massive amounts of data. Existing power grid analysis methods fail to fully consider the dynamic changes in the power grid structure, have low analysis accuracy, and the traditional partitioning method is fixed, making it difficult to adapt to complex power grid operating environments and insufficient optimization capabilities. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] To solve the above technical problems, the present invention provides the following technical solutions: A power grid hierarchical and partitioned intelligent analysis method based on a control cloud comprises the following steps:

[0006] Grid layered and zoned data fusion and feature construction, specifically:

[0007] Under the control cloud architecture, data is collected for power grids at different levels. The short-term characteristics, long-term trend characteristics, and spatial correlation characteristics of power grid data are extracted using spatiotemporal feature extraction functions. Based on the distribution differences of power grid data characteristics at different levels, a hierarchical adaptive clustering algorithm is used to adaptively fuse the data.

[0008] Intelligent analysis and anomaly detection of power grid stratification and zoning, specifically:

[0009] Based on the adaptive fusion results, the hierarchical network structure is used to perform intelligent analysis of the hierarchical partitioning of the power grid, including the hierarchical judgment and partition determination of the power grid data. According to the intelligent analysis results, the state transition matrix is constructed to identify potential abnormalities of the power grid, and also includes:

[0010] By continuously monitoring the abnormal areas in the partition judgment results, the characteristic mean and standard deviation of the power grid at different levels can be dynamically adjusted, thereby achieving dynamic adjustment of the interval variation threshold;

[0011] Intelligent optimization decision-making and control execution of power grid stratification and zoning, specifically:

[0012] Based on the power grid data at different levels, the operating status of each level is analyzed, and intelligent optimization decisions and control execution are made based on abnormal characteristics and global impacts.

[0013] As a preferred solution of the power grid hierarchical and regional intelligent analysis method based on the control cloud of the present invention, the data collection is specifically as follows:

[0014] Based on the control cloud architecture, the operating parameters of different levels of the power grid are collected in real time.

[0015]

[0016] in, Represents the grid level, including, Represents the ultra-high voltage network, Indicates the regional main network and represents the distribution network, Indicates the Collect data at different levels, Indicates the The first level The detection point data vector is a multidimensional data vector, including voltage data, current data, frequency data and power data. Indicates the number of detection points in each layer, and the number of detection points in each layer is the same;

[0017] At the same time, based on the historical database, extract the same historical data set as the real-time power grid data .

[0018] As a preferred solution of the power grid hierarchical and regional intelligent analysis method based on the control cloud of the present invention, the adaptive fusion of data using the hierarchical adaptive clustering algorithm is specifically as follows:

[0019] For the constructed power grid data feature matrix , calculate the characteristic mean of the ultra-high voltage network, regional main network, and distribution network and standard deviation , and determine the data clustering radius at each level , thereby generating the initial cluster center set ;

[0020] For the generated initial cluster center set, calculate the set of data points covered by each cluster center and update the cluster center to generate a new cluster center set ;

[0021] For all cluster centers in the second-generation cluster center set, calculate their density adaptive adjustment factors, and update the positions of the cluster centers based on the adjustment factors to perform hierarchical adaptive clustering. Then,

[0022] Second generation cluster center iteration:

[0023] For the initial cluster center set, by adjusting the cluster center position, ensure that the coverage data of all cluster centers in the adjusted cluster center set is higher than the coverage data of the initial cluster center;

[0024] Three generations of cluster center iteration:

[0025] For all cluster centers in the set of second-generation cluster centers, adjust the positions of the cluster centers to ensure that the consistency score of one cluster center in the set of third-generation cluster centers is higher than that before the adjustment;

[0026] According to the iteration of the second-generation cluster centers and the third-generation cluster centers, until all the second-generation cluster centers can find the third-generation cluster centers with higher cluster consistency scores, the set of all the third-generation cluster centers is the final hierarchical adaptive clustering result. .

[0027] As a preferred solution of the power grid hierarchical and regional intelligent analysis method based on the control cloud of the present invention, the cluster consistency score is specifically as follows:

[0028]

[0029]

[0030] in, represents the total number of cluster centers, represents the category index of the cluster center, Indicates the Cluster centers, Indicates the first cluster center in the second generation cluster center set Cluster centers, Represents the first cluster center in the three generations Cluster centers, Indicates the A set of data points with cluster centers, represents a data point at the cluster center, represents the second generation cluster center set, represents the third generation cluster center set, Represents the cluster consistency score, which is used to implement the iteration of cluster centers. represents the consistency score of the second generation cluster centers, Represents the consistency score of the third-generation cluster centers.

[0031] As a preferred solution of the power grid hierarchical and regional intelligent analysis method based on the control cloud of the present invention, the hierarchical judgment of the power grid data is as follows:

[0032] Input grid data , and make level judgments based on the characteristic mean and standard deviation of different levels of power grids, including the characteristic mean of ultra-high voltage level , regional main network level characteristic mean , distribution network level characteristic mean and, the standard deviation of the ultrahigh pressure level , regional main network level standard deviation , distribution network level standard deviation , then there is,

[0033] If the input power grid data is compared with the characteristic mean and standard deviation of different levels of power grids to meet the formula , indicating that the current input power grid data is at the ultra-high voltage level;

[0034] If the input power grid data is compared with the characteristic mean and standard deviation of different levels of power grids to meet the formula , indicating that the currently input power grid data is at the regional main grid level;

[0035] If the input power grid data is compared with the characteristic mean and standard deviation of different levels of power grids to meet the formula , indicating that the current input power grid data is at the distribution network level.

[0036] As a preferred solution of the power grid hierarchical and regional intelligent analysis method based on the control cloud of the present invention, if the input data meets the judgment conditions of multiple levels at the same time, the adaptability scores of different levels are calculated, and the level judgment is performed according to the calculated adaptability scores, specifically:

[0037] Adaptability Scores at Different Levels , then there is,

[0038]

[0039] in, Indicates the input power grid data, 、 Represent the characteristic mean and standard deviation of different levels of power grid respectively, by controlling The value of controls the selection of different levels of power grids. Represents the adaptability scores of different levels, by calculating the adaptability scores of each level at each level based on the input power grid data, including, 、 as well as , according to the function Conduct hierarchical judgment, including ultra-high voltage network adaptability scoring , Regional main network layer adaptability score and distribution network layer adaptability score , according to the function Make hierarchical judgments.

[0040] As a preferred solution of the power grid hierarchical and partitioned intelligent analysis method based on the control cloud of the present invention, the partition determination is specifically as follows:

[0041] Set stable areas, dynamic adjustment areas, and abnormal areas, and set interval variation threshold ranges , represents the minimum threshold of interval variation, Indicates the maximum threshold of interval variation;

[0042] Calculate the coefficient of variation corresponding to the input power grid data , used for partition determination, specifically:

[0043] If the calculated coefficient of variation satisfies the formula , indicating that the data points in the current level are stable area data;

[0044] If the calculated coefficient of variation satisfies the formula , indicating that the data points in the current level are dynamically adjusted regional data;

[0045] If the calculated coefficient of variation satisfies the formula , indicating that the data points in the current layer are abnormal area data.

[0046] As a preferred solution of the power grid hierarchical and regional intelligent analysis method based on the control cloud of the present invention, the dynamic adjustment of the characteristic mean and standard deviation of the power grids at different levels is as follows:

[0047] Statistics for a time window The total amount of data detected , and collect the number of data points in the abnormal area within the time window , and calculate the anomaly rate within the time window , according to the calculated abnormal rate, the characteristic mean and standard deviation of the power grid are dynamically adjusted, then,

[0048] If the calculated anomaly rate within the time window satisfies the formula , indicating that the data anomaly rate in the current area exceeds the standard. By adjusting the characteristic mean and standard deviation of the power grid, the coefficient of variation can be adjusted, and then the area can be re-determined.

[0049] As a preferred solution of the power grid hierarchical and partitioned intelligent analysis method based on the control cloud of the present invention, the state transition matrix is specifically constructed as follows:

[0050] Statistics for a time window The grid operation status is Duration , and the grid operation status is Duration , and then construct the state transfer matrix, then we have,

[0051]

[0052] in, Indicates the grid operation status is duration, Indicates the grid operation status is duration, Indicates the grid operation status from Transfer to running state The probability of 、 Indicates the power grid operation status category, including stable state, dynamic adjustment state and abnormal state, which correspond to the data judgment area respectively. In the process of constructing the state transfer matrix, 、 The value of must satisfy the formula ;

[0053] According to the calculated state transition probability, the power grid state transition matrix is:

[0054] State transition probability , where the grid operation status category 、 Value, each has three operating states, through 、 The values are different. Based on the state transition probability, a dimension is constructed as The state transition matrix .

[0055] As a preferred solution of the power grid hierarchical and partitioned intelligent analysis method based on the control cloud of the present invention, the identification of potential power grid anomalies by constructing a state transition matrix is specifically as follows:

[0056] Set the state transition probability threshold ,and , based on the set transition probability threshold, the potential abnormality of the power grid is identified, then,

[0057] For any transition probability in the state transition matrix , if the formula is satisfied , indicating that the transfer of the power grid exceeds the upper threshold, the early warning function is immediately triggered. At the same time, the data collection frequency of the current area is increased, and the time series prediction model is used to perform secondary identification of abnormalities. If the secondary identification result is still an abnormal state transfer, it means that the current power grid state is abnormal, and relevant personnel are notified to perform maintenance.

[0058] Beneficial effects of the present invention:

[0059] The present invention integrates power grid data through cloud computing and combines it with intelligent algorithms to perform dynamic hierarchical and partitioned analysis, thus achieving automatic perception and intelligent decision-making of power grid operation status, and improving the automation and intelligence level of power grid regulation.

[0060] By adopting a distributed computing method through a cloud control architecture, efficient storage, calculation, and analysis of data at different levels, such as the ultra-high voltage network, regional main network, and distribution network, are achieved, breaking through the computing power limitations of traditional control models.

[0061] By dynamically adjusting the partition anomaly detection threshold and combining it with global impact analysis to intelligently optimize the anomaly level, it achieves rapid and accurate identification of power grid faults or abnormal conditions, improving the real-time response to power grid anomalies.

[0062] Through dynamic threshold adjustment and state transition matrix analysis based on historical data, trend prediction and precise optimization of power grid operation status are achieved, which reduces misjudgments and unnecessary control operations and improves control reliability.

[0063] Through intelligent hierarchical and regional analysis and regulation, the rational allocation of power grid resources is achieved, manual intervention and energy loss are reduced, the economy of power grid operation is improved, and the overall operation and maintenance costs are reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort. Among them:

[0065] Figure 1 This is a schematic diagram of the overall method steps of the intelligent analysis method for power grid stratification and partitioning based on the control cloud of the present invention. DETAILED DESCRIPTION

[0066] To make the purpose, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0067] Example 1

[0068] Reference Figure 1 , which is the first embodiment of the present invention, provides a power grid hierarchical and partitioned intelligent analysis method based on a control cloud, comprising the following steps:

[0069] S1: Data fusion and feature construction of power grid layering and partitioning.

[0070] Specifically, the grid hierarchical and regional data fusion and feature construction is to collect data from different levels of the grid under the control cloud architecture, and fuse the collected data. At the same time, feature construction is carried out based on the fused grid data to provide an accurate data foundation for subsequent intelligent analysis of grid hierarchical and regional divisions. The specific implementation is as follows:

[0071] Based on the control cloud architecture, the operating parameters of different levels of the power grid are collected in real time.

[0072]

[0073] in, Represents the grid level, including, Represents the ultra-high voltage network, Indicates the regional main network and represents the distribution network, Indicates the Collect data at different levels, Indicates the The first level The detection point data vector is a multidimensional data vector, including voltage data, current data, frequency data and power data. Indicates the number of detection points in each layer, and the number of detection points in each layer is the same;

[0074] At the same time, based on the historical database, extract the same historical data set as the real-time power grid data ,The only difference between the historical data set and the real-time ,collected data is in the time dimension, and the other components are exactly the same;

[0075] According to the real-time collected power grid data and the extracted historical data, the spatiotemporal feature extraction function is used to extract the feature matrix of the power grid data, then,

[0076] The spatiotemporal feature extraction functions are used to extract the short-term features, long-term trend features, and spatial correlation features of the power grid data, specifically:

[0077] Short-term feature extraction,

[0078]

[0079] Long-term trend characteristics,

[0080]

[0081] Spatial correlation characteristics,

[0082]

[0083] Based on the extracted features, the feature matrix of the power grid data is constructed, and then,

[0084]

[0085] in, Indicates the The first level detection point data vectors, Indicates the number of detection points in each layer. The number of detection points in each layer is the same. Represents a collection of historical data, represents the extracted short-term features, represents the extracted spatial correlation features, represents the extracted long-term trend features, represents the short-time feature extraction function, represents the long-term trend feature extraction function, represents the spatial correlation feature extraction function, Represents the power grid data feature matrix;

[0086] Adaptive fusion of data at different levels based on the power grid data matrix, specifically:

[0087] In view of the distribution differences of power grid data characteristics at different levels, based on the hierarchical adaptive clustering algorithm, the adaptive fusion of data at different levels is achieved.

[0088] For the constructed power grid data feature matrix , calculate the characteristic mean of the ultra-high voltage network, regional main network, and distribution network and standard deviation , based on the calculated standard deviation, determine the data clustering radius at each level , thereby generating the initial cluster center set ;

[0089] For the generated initial cluster center set, calculate the set of data points covered by each cluster center, and update the cluster center to generate a new cluster center set (Second generation cluster center), then we have,

[0090] For the initial cluster center set ,in,

[0091]

[0092] in, Indicates the first cluster center in the initial cluster center set cluster centers, Represents the cluster center The set of data points covered, Represents the cluster center Data points covered;

[0093] For the generated second-generation cluster center set, all cluster centers in the second-generation cluster center set are adaptively adjusted, specifically:

[0094] Select all cluster centers in the second-generation cluster center set, calculate their density adaptive adjustment factors, and update the positions of the cluster centers based on the adjustment factors to generate a new cluster center set (third-generation cluster center);

[0095] According to the iteration of the selected initial cluster center set, the second generation cluster center set and the third generation cluster center set, hierarchical adaptive clustering is achieved, then,

[0096] The iteration for the second generation cluster center is:

[0097] For the initial cluster center set, the cluster center position is adjusted by calculating the data point distribution density, requiring that all cluster centers in the adjusted cluster center set can cover a more reasonable data distribution than the initial cluster center;

[0098] The iteration for the three generations of cluster centers is:

[0099] For all cluster centers in the second-generation cluster center set, adjust the position of the cluster center and ensure that at least one cluster center in the third-generation cluster center set can improve the cluster consistency score;

[0100] According to the iteration of the second-generation cluster centers and the third-generation cluster centers, until all the second-generation cluster centers can find the third-generation cluster centers with higher cluster consistency scores, the set of all the third-generation cluster centers is the final hierarchical adaptive clustering result. , specifically:

[0101] Consistency score,

[0102]

[0103]

[0104] in, represents the total number of cluster centers, represents the category index of the cluster center, Indicates the cluster centers, Indicates the A set of data points with cluster centers, represents a data point at the cluster center, represents the second generation cluster center set, represents the third generation cluster center set, Represents the cluster consistency score, which is used to implement the iteration of cluster centers. represents the consistency score of the second generation cluster centers, Represents the consistency score of the third-generation cluster centers.

[0105] S2: Intelligent analysis and anomaly detection of power grid stratification and partitioning.

[0106] Specifically, the intelligent analysis and anomaly detection of power grid hierarchical partitions utilizes a hierarchical network structure and performs intelligent analysis of power grid hierarchical partitions based on the aggregation results of power grid data. Based on the intelligent analysis results, anomaly detection of power grid hierarchical partitions is performed through a dynamic hierarchical architecture. The specific implementation is as follows:

[0107] Aggregation results for power grid data , then there is, , and calculate the characteristic mean and standard deviation of different levels of power grid, then the characteristic mean of ultra-high voltage network is , regional main network characteristic mean and the mean value of distribution network characteristics , standard deviation of the ultra-high voltage network , regional main network standard deviation and the standard deviation of the distribution network ;

[0108] Based on the aggregation results of the power grid data, the hierarchical judgment of the power grid is carried out, specifically:

[0109] Input grid data , and make hierarchical judgments based on the characteristic mean and standard deviation of different levels of power grids, then we have,

[0110] If the input power grid data is compared with the characteristic mean and standard deviation of different levels of power grids to meet the formula , indicating that the current input power grid data is at the ultra-high voltage level;

[0111] If the input grid data meets the formula with the characteristic mean and standard deviation of the grid at different levels, the input grid data meets the formula with the characteristic mean and standard deviation of the grid at different levels. , indicating that the currently input power grid data is at the regional main grid level;

[0112] If the input power grid data is compared with the characteristic mean and standard deviation of different levels of power grids to meet the formula , indicating that the current input power grid data is at the distribution network level.

[0113] It should be noted that if the input data meets the judgment conditions of multiple levels at the same time, the adaptability scores of different levels will be calculated and the level judgment will be made based on the calculated adaptability scores, specifically:

[0114] Adaptability Scores at Different Levels , then there is,

[0115]

[0116] in, Indicates the input power grid data, 、 Represent the characteristic mean and standard deviation of different levels of power grid respectively, by controlling The value of controls the selection of different levels of power grids. Represents the adaptability scores of different levels, by calculating the adaptability scores of each level at each level based on the input power grid data, including, 、 as well as , according to the function Make hierarchical judgments.

[0117] After the initial level determination, the coefficient of variation of the input data is used to determine the partitions, as follows:

[0118] Set stable areas, dynamic adjustment areas, and abnormal areas, and set interval variation threshold ranges ;

[0119] Calculate the coefficient of variation corresponding to the input power grid data, then we have,

[0120]

[0121] in, Indicates the input power grid data, 、 Represent the characteristic mean and standard deviation of different levels of power grid respectively, by controlling The value of controls the selection of different levels of power grids. Indicates the coefficient of variation at different levels, used for partition determination, specifically:

[0122] If the calculated coefficient of variation satisfies the formula , indicating that the data points in the current level are stable area data;

[0123] If the calculated coefficient of variation satisfies the formula , indicating that the data points in the current level are dynamically adjusted regional data;

[0124] If the calculated coefficient of variation satisfies the formula , indicating that the data points in the current layer are abnormal area data.

[0125] It should be noted that in order to improve the accuracy of data partition determination, through continuous monitoring of abnormal areas, the characteristic mean and standard deviation of different levels of power grids are dynamically adjusted, and then the interval variation threshold is dynamically adjusted. Specifically:

[0126] Statistics for a time window The total amount of data detected , and collect the number of data points in the abnormal area within the time window , and calculate the abnormal rate within the time window, then we have,

[0127]

[0128] in, Represents a time window The total amount of data detected within Represents the number of data points in the abnormal area within the time window, It represents the abnormal rate within the time window and is used to control the dynamic adjustment of the characteristic mean and standard deviation of the power grid. Specifically:

[0129] If the calculated anomaly rate within the time window satisfies the formula , indicating that the data anomaly rate in the current area exceeds the standard. By adjusting the characteristic mean and standard deviation of the power grid, the coefficient of variation can be adjusted, and then the area can be re-determined. Then,

[0130]

[0131]

[0132] in, 、 Indicates the adjustment coefficient, which is set by the implementer according to the actual application scenario. 、 Represent the characteristic mean and standard deviation of different levels of power grid, 、 Respectively represent the mean and standard deviation of the historical abnormal data in the abnormal area, 、 Respectively represent the adjusted mean and standard deviation of different levels of power grid;

[0133] The mean and standard deviation of power grids at different levels are adjusted. In order to improve the accuracy of data area determination, the interval variation threshold is adjusted synchronously. Specifically:

[0134]

[0135]

[0136] in, Indicates the adjustment coefficient, which is set by the implementer according to the actual application scenario. 、 Respectively represent the mean and standard deviation of different levels of power grid after adjustment, represents the minimum threshold of interval variation, represents the maximum threshold of interval variation, represents the minimum threshold of interval variation after adjustment, Indicates the maximum threshold of the adjusted interval variation.

[0137] After completing the hierarchical and partitioning determination of the input power grid data, the potential abnormalities of the power grid are identified by constructing a state transition matrix. The specific implementation is as follows:

[0138] Constructing state transition probabilities , indicating that the grid operation status is from Transfer to running state The specific construction process is as follows:

[0139] Statistics for a time window The grid operation status is Duration , and the grid operation status is Duration , and then construct the state transfer matrix, then we have,

[0140]

[0141] in, Indicates the grid operation status is duration, Indicates the grid operation status is duration, Indicates the grid operation status from Transfer to running state The probability of 、 Indicates the power grid operation status category, including stable state, dynamic adjustment state and abnormal state, which correspond to the data judgment area respectively. In the process of constructing the state transfer matrix, 、 The value of must satisfy the formula ;

[0142] According to the calculated state transition probability, the power grid state transition matrix is:

[0143] State transition probability , where the grid operation status category 、 Value, each has three operating states, so through 、 The values are different. Based on the state transition probability, a dimension is constructed as The state transition matrix , then there is,

[0144]

[0145] in, represents the constructed state transition matrix, Indicates the grid operation status from Transfer to running state The probability of 、 Indicates the grid operation status category;

[0146] Identify potential power grid anomalies based on the constructed state transition matrix, specifically:

[0147] Set the state transition probability threshold ,and , based on the set transition probability threshold, the potential abnormality of the power grid is identified, then,

[0148] For any transition probability in the state transition matrix , if the formula is satisfied , indicating that the transfer of the power grid exceeds the upper threshold, the early warning function is immediately triggered. At the same time, the data collection frequency of the current area is increased, and the time series prediction model is used to perform secondary identification of abnormalities. If the secondary identification result is still an abnormal state transfer, it means that the current power grid state is abnormal, and relevant personnel are notified to perform maintenance.

[0149] S3: Intelligent optimization decision-making and control execution of power grid layering and partitioning.

[0150] Specifically, the grid hierarchical and regional intelligent optimization decision-making and control execution is based on the grid data at different levels, analyzing the operating status of each level, and making intelligent optimization decisions and control execution based on abnormal characteristics and global impacts, as follows:

[0151] Based on the grid status at different levels, intelligent optimization of layers and zones is carried out, specifically:

[0152] Based on the input power grid data, the mean and standard deviation of the operating parameters of the ultra-high voltage network, regional main network, and distribution network are calculated, and the abnormality judgment threshold is set based on the standard deviation to preliminarily screen abnormal areas;

[0153] For the abnormal areas screened out, the status of adjacent areas is analyzed to determine whether the abnormality has a global impact, and the abnormality level is adjusted accordingly to generate a basis for optimized decision-making;

[0154] Based on the optimization decision basis, intelligent control is performed. Specifically, the following steps are performed: select areas with persistent abnormalities, calculate their load adjustment factors and voltage optimization factors, and adjust operating parameters based on these factors to optimize the regional status;

[0155] According to the abnormal areas initially screened, the basis for optimization decision-making and the iteration of intelligent control execution, hierarchical and partitioned intelligent optimization is achieved.

[0156] The iteration for optimizing the decision basis is:

[0157] For the abnormal areas initially screened, the abnormality level is adjusted through global impact analysis, and the adjusted abnormality level is required to more accurately reflect the operating status of the power grid;

[0158] The iterations executed for intelligent control are:

[0159] Adjust operating parameters for all abnormal areas in the optimization decision basis and ensure that at least one area can be restored to a stable state to improve the overall stability of the power grid;

[0160] Based on the optimization decision basis and the iteration of intelligent control execution, until reasonable optimization adjustment solutions can be found for all areas with persistent abnormalities, the set of all optimized areas is the final hierarchical and partitioned intelligent optimization decision and control execution result.

[0161] Furthermore, if the functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or the portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage media include various media that can store program code, such as USB flash drives, mobile hard drives, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical disks.

[0162] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0163] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, and then editing, interpreting, or processing in another suitable manner as necessary, and then storing it in a computer memory.

[0164] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the protection scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A cloud-based intelligent analysis method for power grid stratification and partitioning, characterized by: The following steps are included: Grid layered and zoned data fusion and feature construction, specifically: Under the control cloud architecture, data is collected for power grids at different levels. The short-term characteristics, long-term trend characteristics, and spatial correlation characteristics of power grid data are extracted using spatiotemporal feature extraction functions. Based on the distribution differences of power grid data characteristics at different levels, a hierarchical adaptive clustering algorithm is used to adaptively fuse the data. Intelligent analysis and anomaly detection of power grid stratification and zoning, specifically: Based on the adaptive fusion results, the hierarchical network structure is used to perform intelligent analysis of the hierarchical partitioning of the power grid, including the hierarchical judgment and partition determination of the power grid data. According to the intelligent analysis results, the state transition matrix is constructed to identify potential abnormalities of the power grid, and also includes: By continuously monitoring the abnormal areas in the partition judgment results, the characteristic mean and standard deviation of the power grid at different levels can be dynamically adjusted, thereby achieving dynamic adjustment of the interval variation threshold; Intelligent optimization decision-making and control execution of power grid stratification and zoning, specifically: Analyze the operating status of each level based on grid data at different levels, and make intelligent optimization decisions and control execution based on abnormal characteristics and global impacts; The adaptive fusion of data using the hierarchical adaptive clustering algorithm is specifically as follows: For the constructed power grid data feature matrix , calculate the characteristic mean of the ultra-high voltage network, regional main network, and distribution network and standard deviation , and determine the data clustering radius at each level , thereby generating the initial cluster center set ; For the generated initial cluster center set, calculate the set of data points covered by each cluster center and update the cluster center to generate a new cluster center set ; For all cluster centers in the second-generation cluster center set, calculate their density adaptive adjustment factors, and update the positions of the cluster centers based on the adjustment factors to perform hierarchical adaptive clustering. Then, Second generation cluster center iteration: For the initial cluster center set, by adjusting the cluster center position, ensure that the coverage data of all cluster centers in the adjusted cluster center set is higher than the coverage data of the initial cluster center; Three generations of cluster center iteration: For all cluster centers in the set of second-generation cluster centers, adjust the positions of the cluster centers to ensure that the consistency score of one cluster center in the set of third-generation cluster centers is higher than that before the adjustment; According to the iteration of the second-generation cluster centers and the third-generation cluster centers, until all the second-generation cluster centers can find the third-generation cluster centers with higher cluster consistency scores, the set of all the third-generation cluster centers is the final hierarchical adaptive clustering result. .

2. The power grid hierarchical and regional intelligent analysis method based on the control cloud according to claim 1 is characterized in that: The data collection is specifically as follows: Based on the control cloud architecture, the operating parameters of different levels of the power grid are collected in real time. in, Represents the grid level, including, Represents the ultra-high voltage network, Indicates the regional main network and represents the distribution network, Indicates the Collect data at different levels, Indicates the The first level The detection point data vector is a multidimensional data vector, including voltage data, current data, frequency data and power data. Indicates the number of detection points in each layer, and the number of detection points in each layer is the same; At the same time, based on the historical database, extract the same historical data set as the real-time power grid data .

3. The method for intelligent analysis of power grid hierarchical zoning based on control cloud according to claim 2 is characterized in that: The cluster consistency scores are as follows: in, represents the total number of cluster centers, represents the category index of the cluster center, Indicates the Cluster centers, Indicates the first cluster center in the second generation cluster center set Cluster centers, Represents the first cluster center in the three generations Cluster centers, Indicates the A set of data points with cluster centers, represents a data point at the cluster center, represents the second generation cluster center set, represents the third generation cluster center set, Represents the cluster consistency score, which is used to implement the iteration of cluster centers. represents the consistency score of the second generation cluster centers, Represents the consistency score of the third-generation cluster centers.

4. The power grid hierarchical and regional intelligent analysis method based on the control cloud according to claim 3 is characterized in that: The hierarchy of the power grid data is specifically determined as follows: Input grid data , and make level judgments based on the characteristic mean and standard deviation of different levels of power grids, including the characteristic mean of ultra-high voltage level , regional main network level characteristic mean , distribution network level characteristic mean and, the standard deviation of the ultrahigh pressure level , regional main network level standard deviation , distribution network level standard deviation , then there is, If the input power grid data is compared with the characteristic mean and standard deviation of different levels of power grids to meet the formula , indicating that the current input power grid data is at the ultra-high voltage level; If the input power grid data is compared with the characteristic mean and standard deviation of different levels of power grids to meet the formula , indicating that the currently input power grid data is at the regional main grid level; If the input power grid data is compared with the characteristic mean and standard deviation of different levels of power grids to meet the formula , indicating that the current input power grid data is at the distribution network level.

5. The method for intelligent analysis of power grid hierarchical zoning based on control cloud according to claim 4 is characterized in that: If the input data meets the judgment conditions of multiple levels at the same time, the adaptability scores of different levels are calculated and the level judgment is performed based on the calculated adaptability scores. Specifically: Adaptability scores at different levels , then there is, in, Indicates the input power grid data, 、 Represent the characteristic mean and standard deviation of different levels of power grid respectively, by controlling The value of controls the selection of different levels of power grids. Indicates the adaptability scores of different levels. By calculating the input grid data, the adaptability scores at each level, including the adaptability scores of the ultra-high voltage grid layer , Regional main network layer adaptability score and distribution network layer adaptability score , according to the function Make hierarchical judgments.

6. The method for intelligent analysis of power grid hierarchical zoning based on control cloud according to claim 5 is characterized in that: The partition determination is specifically as follows: Set stable areas, dynamic adjustment areas, and abnormal areas, and set interval variation threshold ranges , represents the minimum threshold of interval variation, Indicates the maximum threshold of interval variation; Calculate the coefficient of variation corresponding to the input power grid data , used for partition determination, specifically: If the calculated coefficient of variation satisfies the formula , indicating that the data points in the current level are stable area data; If the calculated coefficient of variation satisfies the formula , indicating that the data points in the current level are dynamically adjusted regional data; If the calculated coefficient of variation satisfies the formula , indicating that the data points in the current layer are abnormal area data.

7. The method for intelligent analysis of power grid hierarchical zoning based on control cloud according to claim 6, characterized in that: The dynamic adjustment of the characteristic mean and standard deviation of the power grids at different levels is as follows: Statistics for a time window The total amount of data detected , and collect the number of data points in the abnormal area within the time window , and calculate the anomaly rate within the time window , according to the calculated abnormal rate, the characteristic mean and standard deviation of the power grid are dynamically adjusted, then, If the calculated anomaly rate within the time window satisfies the formula , indicating that the data anomaly rate in the current area exceeds the standard. By adjusting the characteristic mean and standard deviation of the power grid, the coefficient of variation can be adjusted, and then the area can be re-determined.

8. The method for intelligent analysis of power grid hierarchical zoning based on control cloud according to claim 7, characterized in that: The state transfer matrix is specifically constructed as follows: Statistics for a time window The grid operation status is Duration , and the grid operation status is Duration , and then construct the state transfer matrix, then we have, in, Indicates the grid operation status is duration, Indicates the grid operation status is duration, Indicates the grid operation status from Transfer to running state The probability of 、 Indicates the power grid operation status category, including stable state, dynamic adjustment state and abnormal state, which correspond to the data judgment area respectively. In the process of constructing the state transfer matrix, 、 The value of must satisfy the formula ; According to the calculated state transition probability, the power grid state transition matrix is: State transition probability , where the grid operation status category 、 Value, each has three operating states, through 、 The values are different. Based on the state transition probability, a dimension is constructed as The state transition matrix .

9. The method for intelligent analysis of power grid hierarchical zoning based on control cloud according to claim 8, characterized in that: The identification of potential abnormalities in the power grid by constructing a state transition matrix is specifically as follows: Set the state transition probability threshold ,and , based on the set transition probability threshold, the potential abnormality of the power grid is identified, then, For any transition probability in the state transition matrix , if the formula is satisfied , indicating that the transfer of the power grid exceeds the upper threshold, the early warning function is immediately triggered. At the same time, the data collection frequency of the current area is increased, and the time series prediction model is used to perform secondary identification of abnormalities. If the secondary identification result is still an abnormal state transfer, it means that the current power grid state is abnormal, and relevant personnel are notified to perform maintenance.

Images