Dynamic construction method of electrical equipment digital twin suitable for power distribution automation

Abstract

The application discloses a kind of electrical equipment digital twin dynamic construction methods suitable for power distribution automation, it is related to digital twin system technical field, can solve how to guarantee the premise of effective monitoring and management to lighting equipment in distribution area, dynamically determine the reasonable construction granularity and precision of digital twin problem, including: obtaining current data in target area;According to the distribution characteristic index of the lighting device contained in different lighting lines in target area determined by current data, and accordingly, all lighting lines are clustered and divided into multiple to-be-allocated line clusters;According to the connection relationship between each to-be-allocated line cluster inside each lighting line and the connection relationship of lighting line between different to-be-allocated line clusters, the modified electric energy scene characteristic index is determined;According to distribution characteristic index and modified electric energy scene characteristic index, all lighting lines are re-clustered, form multiple standard allocation line cluster sets and dynamically construct digital twin for each set.

Description

Technical Field

[0001] This invention relates to the field of digital twin system technology, and specifically to a method for dynamically constructing digital twins of electrical equipment suitable for power distribution automation. Background Technology

[0002] With the deepening application of digital twin technology in the industrial field, its prospects for improving the management efficiency and intelligent operation and maintenance of electrical equipment in power distribution automation systems are attracting increasing attention. By constructing a virtual mapping of physical entities, digital twins can achieve real-time monitoring, simulation analysis, and predictive maintenance of equipment status, providing a new technological path for transforming the traditional passive power distribution management model that relies on manual inspections. Especially in urban-level power distribution networks, which connect massive amounts of terminal electrical equipment such as lighting devices, the refined perception and collaborative optimization of their operating status is one of the key links in realizing power distribution automation.

[0003] Currently, in power distribution automation applications such as urban lighting, which involve large-scale, widely distributed, and diverse equipment in various scenarios, efficiently constructing and managing corresponding digital twins has become an important area of ​​technological exploration. Existing digital twin construction methods typically tend to create independent or generalized virtual models for physical entities. However, when faced with a massive number of urban lighting devices with diverse scenario requirements, simply constructing a high-precision twin for each device would lead to a sharp increase in the cost of model construction, data computation, and synchronous maintenance, resulting in enormous resource consumption. On the other hand, using overly coarse aggregated models would be insufficient to support the need for precise perception and optimized control of device operating status under differentiated scenarios. Therefore, effectively balancing the model accuracy, management granularity, and implementation cost of digital twins while meeting the requirements of power distribution automation management has become a common technical challenge that urgently needs to be addressed. Summary of the Invention

[0004] To address the current technical challenge of dynamically determining the appropriate granularity and precision of digital twin construction while ensuring effective monitoring and management of lighting equipment within a power distribution area, thereby optimizing overall construction costs and resource consumption, this invention aims to provide a method for dynamically constructing digital twins of electrical equipment suitable for power distribution automation. The specific technical solution adopted is as follows: In a first aspect, the present invention provides a method for dynamically constructing digital twins of electrical equipment applicable to power distribution automation, comprising: acquiring current data within a target area; wherein the current data includes line current data of each lighting line within the target area and device current data of the corresponding lighting device; determining the distribution characteristic index of the lighting devices contained in different lighting lines within the target area based on the current data, and clustering all lighting lines into multiple unassigned line clusters based on the distribution characteristic index; determining the power scene characteristic index of each lighting line based on the connection relationship between each lighting line within each unassigned line cluster; correcting the power scene characteristic index based on the connection relationship between lighting lines in different unassigned line clusters to obtain corrected power scene characteristic index; re-clustering all lighting lines based on the distribution characteristic index and the corrected power scene characteristic index to form multiple standard assigned line cluster sets, and dynamically constructing a digital twin for each standard assigned line cluster set.

[0005] Secondly, this invention provides a dynamic construction system for digital twins of electrical equipment suitable for power distribution automation, comprising: a data acquisition module, a line clustering module, a scene feature determination module, a line re-clustering module, and a digital twin construction module; the data acquisition module is used to acquire current data within a target area; wherein, the current data includes line current data of each lighting line within the target area and device current data of the corresponding lighting device; the line clustering module is used to determine the distribution characteristic index of the lighting devices contained in different lighting lines within the target area based on the current data, and to cluster all lighting lines into multiple unassigned line clusters based on the distribution characteristic index; the scene feature determination module is used to determine the power scene characteristic index of each lighting line based on the connection relationship between each lighting line within each unassigned line cluster; it is also used to correct the power scene characteristic index based on the connection relationship between lighting lines in different unassigned line clusters to obtain corrected power scene characteristic index; the line re-clustering module is used to re-cluster all lighting lines based on the distribution characteristic index and the corrected power scene characteristic index to form multiple standard assigned line cluster sets; the digital twin construction module is used to dynamically construct a digital twin for each standard assigned line cluster set.

[0006] Thirdly, the present invention provides an electronic device, comprising: a processor and a memory; wherein the memory is used to store one or more programs, the one or more programs including computer-executable instructions, and when the electronic device is running, the processor executes the computer-executable instructions stored in the memory to cause the electronic device to perform the dynamic construction method for digital twins of electrical equipment suitable for power distribution automation as described in the first aspect and any possible implementation thereof.

[0007] This invention offers the following advantages: By acquiring current data from lighting lines and devices within a power distribution area, it sequentially performs initial clustering based on device distribution characteristics, scene feature quantification based on intra-cluster connectivity, feature correction based on inter-cluster connectivity, and secondary clustering based on dual features. Finally, it constructs a digital twin based on the importance of the clustering results. This method creatively combines the physical distribution, topological connections, and operational characteristics of power lines, dynamically and intelligently determining the appropriate granularity and precision for digital twin construction. This effectively optimizes the overall construction cost and resource utilization efficiency of the digital twin system while meeting the needs of power distribution automation for equipment status monitoring and management. Attached Figure Description

[0008] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 This is a schematic diagram of the architecture of a dynamic construction system for digital twins of electrical equipment suitable for power distribution automation, provided in one embodiment of the present invention. Figure 2 This is a flowchart illustrating a method for dynamically constructing digital twins of electrical equipment suitable for power distribution automation, as provided in one embodiment of the present invention. Detailed Implementation

[0010] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the specific implementation methods, structures, features, and effects of the present invention will be described in detail below with reference to the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0011] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0012] In all division and logarithmic operations involved in this invention, a smoothing mechanism is employed to prevent computer program crashes or invalid values ​​from being generated due to a zero denominator or zero input. Specifically, a correction factor ε, which is a very small positive number, is superimposed on the denominator term of the division operation or the argument term of the logarithmic function, for example, a value of 10 to the power of negative 5, thereby ensuring the robustness and feasibility of the algorithm under extreme conditions.

[0013] Unless otherwise specified, the normalization function Norm() mentioned in this invention uses maximum and minimum value normalization. The maximum and minimum values ​​are preset empirical extreme values ​​derived from a large amount of historical experimental data. If the calculation result exceeds the [0, 1] interval, it is restricted to the [0, 1] range by a truncation function (i.e., if the result is less than 0, it is taken as 0; if it is greater than 1, it is taken as 1) to eliminate the influence of outliers on the evaluation index.

[0014] The specific scheme of the dynamic construction method for digital twins of electrical equipment applicable to power distribution automation provided by the present invention will be described in detail below with reference to the accompanying drawings.

[0015] For example, such as Figure 1 The diagram shown is an architectural schematic of a dynamic construction system for digital twins of electrical equipment suitable for power distribution automation (hereinafter referred to as the dynamic construction system) provided in an embodiment of the present invention. The dynamic construction system 10 includes: a data acquisition module 11, a line clustering module 12, a scene feature determination module 13, a line re-clustering module 14, and a digital twin construction module 15. The modules are described in detail below: (1) Data acquisition module 11.

[0016] The data acquisition module 11 is responsible for collecting basic current data required to construct a digital twin from the power distribution automation system, smart meters and various sensing devices in the target area, providing raw data input for subsequent analysis and clustering.

[0017] Optionally, the data acquisition module 11 is used to acquire current data within the target area. Specifically, the current data includes: line current data of each lighting circuit within the target area, and device current data of each lighting device connected to these circuits.

[0018] In practice, the data acquisition module 11 first divides the vast target area into several more easily processed sub-regions based on preset regional division rules (such as administrative jurisdictions or power grids). Then, within each sub-region, the data acquisition module 11 synchronously collects line current data from line monitoring points and device current data from smart lighting devices at a preset sampling frequency (e.g., once per second) via an IoT gateway or data concentrator. For conventional lighting devices, their device current data can be indirectly correlated or estimated through monitoring point data from their respective lines. Finally, the data acquisition module 11 performs time-stamp alignment, format standardization, and preliminary verification on the collected raw data to form a well-organized current data set, which is then output to the line clustering module 12.

[0019] (2) Line clustering module 12.

[0020] The line clustering module 12 is responsible for receiving current data from the data acquisition module 11. By analyzing the composition and distribution of lighting devices in different lighting lines, it calculates distribution characteristic indicators that can characterize the density of distribution and scene specificity, and performs initial clustering of all lines based on these indicators to form multiple line clusters to be assigned.

[0021] Optionally, the line clustering module 12 is used to determine the distribution characteristic index of the lighting devices contained in different lighting lines within the target area based on the current data, and to cluster all lighting lines into multiple line clusters to be assigned based on the distribution characteristic index.

[0022] In practice, the line clustering module 12 first analyzes the current data to identify intelligent lighting devices (capable of detecting their own current data) and conventional lighting devices (not possessing the aforementioned capability) in each lighting line. Then, the line clustering module 12 classifies lines containing at least one intelligent lighting device as intelligent lighting lines and lines without intelligent lighting devices as conventional lighting lines.

[0023] For each smart lighting line, the line clustering module 12 calculates and normalizes its first distribution characteristic index based on the first number of smart lighting devices contained in the line, the total number of smart lighting devices contained in all smart lighting lines, and the positional distance between the smart lighting devices in the line. For each conventional lighting line, the line clustering module 12 directly uses the second number of conventional lighting devices it contains as its second distribution characteristic index.

[0024] Finally, the line clustering module 12 performs a first clustering of all smart lighting lines based on the differences between the first distribution characteristic indicators of all smart lighting lines using a preset clustering algorithm (such as DBSCAN), forming at least one first-class line cluster to be assigned. Simultaneously, based on the differences between the second distribution characteristic indicators (i.e., the number of conventional lighting devices) of all conventional lighting lines, it performs a second clustering using the same clustering algorithm, forming at least one second-class line cluster to be assigned. These line clusters to be assigned output by the line clustering module 12 will serve as the analysis units of the scene feature determination module 13.

[0025] (3) Scene feature determination module 13.

[0026] The scene feature determination module 13 is responsible for performing two core calculations in sequence: First, it analyzes the connection topology of the lighting circuits within each cluster of circuits to be assigned and determines the initial power scene feature index of each circuit; then, it further analyzes the connection relationship of the lighting circuits between different clusters of circuits to be assigned, corrects the initial index, and outputs the final corrected power scene feature index.

[0027] Optionally, the scene feature determination module 13 is used to determine the power scene feature index of each lighting line based on the connection relationship between each lighting line within each cluster to be assigned; and to correct the power scene feature index based on the connection relationship between lighting lines in different clusters to be assigned, so as to obtain the corrected power scene feature index.

[0028] For example, the scene feature determination module 13 can be divided into an internal feature calculation submodule 131 and a cross-cluster correction submodule 132, to respectively complete intra-cluster feature quantization and inter-cluster influence correction, which will be described below: (3.1) Internal feature calculation submodule 131.

[0029] Optionally, the internal feature calculation submodule 131 is used to determine the power scene feature indicators of each lighting line based on the connection relationship between each lighting line within each cluster of lines to be assigned.

[0030] Specifically, the internal feature calculation submodule 131 first receives the line cluster data to be assigned from the line clustering module 12. For any current lighting line, the internal feature calculation submodule 131 identifies other lighting lines directly connected to it from the network topology as overlapping lighting lines and locates the connection point, i.e., the overlapping node.

[0031] Subsequently, the internal feature calculation submodule 131 obtains the lighting devices located upstream of the overlapping nodes in the current lighting line to form a sequence of overlapping lighting devices, and calculates the line reuse influence factor of the overlapping lighting line based on the number of lighting devices contained in the sequence and the amount of current reduced after the current lighting line flows through each node.

[0032] Simultaneously, it combines the average current variation, the average current variance of the overlapping lighting device sequence and the overlapping lighting line itself, and the positional distance between devices on both sides of the overlapping node to determine the power scene characteristic indicators of the current lighting line through comprehensive calculation. This process generates an initial characteristic quantification value for each connected line within the cluster.

[0033] (3.2) Cross-cluster correction submodule 132.

[0034] Optionally, the cross-cluster correction submodule 132 is used to correct the power scene characteristic indicators according to the connection relationship of lighting lines between different line clusters to be assigned, so as to obtain the corrected power scene characteristic indicators.

[0035] Specifically, the cross-cluster correction submodule 132 receives initial power scene feature indicators from the internal feature calculation submodule 131. For the current lighting line to be corrected, the cross-cluster correction submodule 132 first scans all lighting lines in all line clusters to be assigned and marks them as cross-cluster connected lines.

[0036] Then, the cross-cluster correction submodule 132 filters out lines from these cross-cluster connection lines that do not belong to the line cluster to which the current lighting line belongs, defines them as reference lighting lines, and defines the other line clusters to which these reference lighting lines belong as reference clusters.

[0037] Finally, the cross-cluster correction submodule 132 adjusts the initial power scene characteristic index of the current lighting line by weighting or proportionally based on the number of identified reference lighting lines and the number of reference clusters, and calculates the final corrected power scene characteristic index that can reflect the cross-cluster coupling effect.

[0038] The corrected power scene feature index output by the scene feature determination module 13 will be input into the line re-clustering module 14 together with the distribution feature index generated by the line clustering module 12.

[0039] (4) Line re-clustering module 14.

[0040] The line re-clustering module 14 is responsible for receiving the distribution characteristic indicators from the line clustering module 12 and the corrected power scene characteristic indicators from the scene feature determination module 13. Using these two sets of indicators as joint features, it performs a more refined and comprehensive clustering analysis on all lighting lines to form an optimized set of logical units that are ultimately used for the construction of the digital twin, namely the standard allocation line cluster set.

[0041] Optionally, the line re-clustering module 14 is used to re-cluster all lighting lines according to the distribution characteristic index and the modified power scene characteristic index to form multiple standard allocation line cluster sets.

[0042] In specific implementation, the line re-clustering module 14 first processes the smart lighting lines: it calculates the difference degree of the first distribution characteristic index between each pair of smart lighting lines, and the difference degree of their corrected power scene characteristic index; then, it merges the two difference degrees (e.g., takes the average value) to obtain the first comprehensive difference degree used to measure the comprehensive difference between lines; finally, based on the first comprehensive difference degree matrix, it performs a third clustering on all smart lighting lines using a preset clustering algorithm to form at least one smart standard cluster.

[0043] For conventional lighting lines, the line re-clustering module 14 adopts the same process: calculating the difference between the second distribution characteristic index (i.e., the number of conventional lighting devices) and the difference between the corrected power scene characteristic index between each pair of lines, fusing them to obtain the second comprehensive difference, and performing the fourth clustering based on this to form at least one conventional standard cluster. All generated intelligent standard clusters and conventional standard clusters together constitute the final set of standard allocation line clusters.

[0044] The line re-clustering module 14 outputs this result to the digital twin construction module 15 as the target object for twin construction.

[0045] (5) Digital twin construction module 15.

[0046] The digital twin construction module 15 is responsible for receiving the standard allocation line cluster set output by the line re-clustering module 14, and dynamically constructing digital twins with different precision and update frequencies for each set based on the comprehensive feature evaluation results, thereby realizing fine management of power distribution automation under resource optimization.

[0047] Optionally, the digital twin building module 15 is used to dynamically build a digital twin for each standard-assigned set of line clusters.

[0048] In its implementation, the digital twin construction module 15 first performs an evaluation decision: it analyzes the distribution characteristics and modified power scenario characteristics of all lighting lines within each standard allocation line cluster set, and determines a comprehensive priority that characterizes the management priority of the set through statistical and fusion calculations (such as weighted averaging). Then, the digital twin construction module 15 compares the comprehensive priority of each set with the system's preset priority threshold.

[0049] For sets with a comprehensive priority greater than a preset threshold, the digital twin construction module 15 calls the high-precision modeling engine and real-time data synchronization service to construct a digital twin with high geometric accuracy and high refresh rate to support second-level monitoring and deep simulation.

[0050] For sets with a comprehensive priority less than or equal to a preset threshold, the digital twin construction module 15 invokes lightweight modeling tools and low-frequency data update services to construct digital twins with lower geometric accuracy and lower update frequency, meeting basic status monitoring requirements. Through this differentiated construction strategy, the system achieves an optimal balance between digital twin construction cost, operational load, and management efficiency.

[0051] The above describes the functions of the dynamic construction system 10 and its modules.

[0052] For example, such as Figure 2 The diagram shown is a flowchart illustrating a method for dynamically constructing digital twins of electrical equipment suitable for power distribution automation, according to an embodiment of the present invention. The method includes the following steps: S201. Obtain current data within the target area. This current data includes the line current data of each lighting circuit within the target area and the device current data of the corresponding lighting fixture.

[0053] For example, this step can be performed by the data acquisition module 11 in the dynamic construction system 10 described above.

[0054] Specifically, the data acquisition module 11 first divides the large target area into multiple independent analysis sub-regions according to preset regional division rules (such as using municipal districts or power grids as boundaries) to reduce the amount of data and computational complexity of a single data processing.

[0055] Subsequently, within each analysis sub-region, the data acquisition module 11 continuously collects line current data for each lighting line at a preset sampling frequency (e.g., once per second) through monitoring terminals deployed at the starting end and key electrical topology nodes (such as branch points and load connection points) of each lighting line. Simultaneously, for intelligent lighting devices with self-detection capabilities, the data acquisition module 11 directly collects device current data through their built-in sensors; for conventional lighting devices without self-detection capabilities, their device current data can be calculated using the current difference between the upstream and downstream monitoring points of the corresponding line at the same time. This difference calculation method is a conventional existing technology in the field of power distribution monitoring. The data acquisition module 11 performs time-scale alignment, format standardization, and preliminary outlier filtering on the collected multi-source current data to form a well-organized current dataset suitable for subsequent analysis.

[0056] Therefore, the data acquisition module 11 provides accurate, synchronous and structured basic data input for the entire dynamic construction process by dividing the target area into sub-regions and collecting multi-source current data at a fixed frequency.

[0057] S202. Determine the distribution characteristics of lighting devices in different lighting circuits within the target area based on the current data, and cluster all lighting circuits into multiple circuit clusters to be assigned based on the distribution characteristics.

[0058] For example, this step can be executed by the line clustering module 12 in the dynamically constructed system 10 described above. Specifically, the line clustering module 12 first identifies and distinguishes between intelligent lighting devices and conventional lighting devices, and then classifies the lines into intelligent lighting lines and conventional lighting lines. For intelligent lighting lines, the line clustering module 12 calculates a first distribution characteristic index by combining the number of intelligent lighting devices within the line, the total number of intelligent lighting devices globally, and the spatial relationship between the devices. For conventional lighting lines, the line clustering module 12 directly uses the number of conventional lighting devices it contains as a second distribution characteristic index. Finally, the line clustering module 12 clusters the two types of lines using a preset clustering algorithm based on the differences between the first distribution characteristic index and the differences between the second distribution characteristic index, forming multiple line clusters to be assigned. It should be noted that the specific process of the aforementioned sub-steps is described in S301-S305 below, and will not be repeated here.

[0059] In another possible implementation, when determining the distribution characteristic index of lighting devices contained in different lighting lines within the target area, the line clustering module 12 can also, based on the line classification, not only consider the number and distribution of lighting devices, but also introduce the historical average load rate of the line during typical periods as a correction factor to jointly calculate a composite distribution characteristic index that reflects the "functional density" of the line.

[0060] In another possible implementation, when the line clustering module 12 divides all lighting lines into multiple line clusters to be assigned, it can first use a hierarchical clustering algorithm to generate preliminary clustering results, and then automatically select the number of clusters and the division scheme corresponding to the optimal profile coefficient by calculating the profile coefficient under different numbers of clusters, thereby determining the final line clusters to be assigned, so as to improve the objectivity and internal consistency of the clustering results.

[0061] Therefore, the line clustering module 12 quantitatively analyzes the composition and spatial distribution characteristics of lighting devices in the line, and performs preliminary clustering based on this, organizing the massive and scattered lighting lines into several groups with relatively consistent internal characteristics to be analyzed, laying a structural foundation for subsequent refined scene feature analysis.

[0062] S203. Based on the connection relationship between each lighting circuit within each cluster of circuits to be assigned, determine the power scene characteristic indicators of each lighting circuit.

[0063] For example, this step can be performed by the scene feature determination module 13 in the dynamic construction system 10 described above. Specifically, the scene feature determination module 13 analyzes the physical connection topology between the lighting lines within each cluster of lines to be assigned. For any current lighting line, it identifies the overlapping lighting lines and connection points (overlapping nodes) directly connected to it. Subsequently, it obtains the lighting devices located upstream of the overlapping nodes in the current lighting line to form a sequence of overlapping lighting devices. Then, combining the composition information of the overlapping lighting device sequence, current shunting data, and the current fluctuation characteristics of the overlapping lighting line itself, it determines the power scene feature index of each overlapping lighting line through comprehensive calculation. This index quantifies the power demand and fluctuation characteristics exhibited by the line under a specific internal connection relationship. It should be noted that the specific process of the aforementioned sub-steps is described in S401-S403 below, and will not be repeated here.

[0064] In another possible implementation, when determining the power scene characteristic indicators of each lighting circuit, the scene feature determination module 13 may not simply use the arithmetic mean when calculating the mean and variance of current change, but instead assign different weights based on the data point collection time (such as higher weight for recent data), thereby calculating the weighted average mean of current change and the weighted average of variance, so that the final determined power scene characteristic indicators can better reflect the recent operating status of the system.

[0065] Therefore, the scenario feature determination module 13 transforms qualitative line connections into quantitative power scenario feature indicators by deeply analyzing the topological connections and power interaction relationships within the cluster of lines to be allocated. This achieves the mapping from physical topology to functional features and provides data basis for evaluating the personalized operation requirements of the lines.

[0066] S204. Based on the connection relationship of lighting circuits between different clusters of circuits to be allocated, the power scene characteristic indicators are corrected to obtain the corrected power scene characteristic indicators.

[0067] For example, this step can be performed by the scene feature determination module 13 in the dynamic construction system 10 described above. Specifically, after obtaining the initial power scene feature indicators of each line, the scene feature determination module 13 further analyzes the cross-connection relationships between different line clusters to be assigned. For the current lighting line, all lighting lines directly connected to it but belonging to other line clusters to be assigned (i.e., cross-cluster connection lines) are scanned and defined as reference lighting lines, and their respective clusters are defined as reference clusters. Finally, based on the number of identified reference lighting lines and reference clusters, the initial power scene feature indicators of the current lighting line are proportionally adjusted or numerically superimposed to obtain corrected power scene feature indicators. This correction process ensures that the indicators can reflect the power coupling effect across clusters. It should be noted that the specific process of the aforementioned sub-steps is described in S501-S503 below, and will not be repeated here.

[0068] In another possible implementation, when correcting the power scene characteristic indicators, the scene feature determination module 13 can not only consider the existence of reference lighting lines and reference clusters, but also further consider the electrical distance (such as line impedance value or shortest path hop count) between the current lighting line and each reference lighting line. The magnitude of the correction is positively correlated with the number of reference objects, but negatively correlated with the average electrical distance, thus more accurately characterizing the intensity of cross-cluster influence.

[0069] Therefore, the scene feature determination module 13 makes important supplementary corrections to the initial power scene feature indicators by identifying and quantifying the topological coupling effect between the line clusters to be allocated, so that the final corrected power scene feature indicators can comprehensively and accurately characterize the overall scene features and demand intensity of each lighting line in the overall power distribution network.

[0070] S205. Based on the distribution characteristic index and the modified power scenario characteristic index, all lighting lines are re-clustered to form multiple standard allocation line cluster sets, and a digital twin is dynamically constructed for each standard allocation line cluster set.

[0071] For example, this step can be executed by the line re-clustering module 14 and the digital twin construction module 15 in the dynamic construction system 10 described above. Specifically, it includes: First, the line re-clustering module 14 uses the distribution characteristic index (first or second distribution characteristic index) and the modified power scene characteristic index of each lighting line as a joint feature vector. For smart lighting lines, it calculates the comprehensive value (first comprehensive difference) of the difference between the above two indicators between each pair of lines and performs clustering to form smart standard clusters; for conventional lighting lines, it uses the same principle to form conventional standard clusters. All smart standard clusters and conventional standard clusters together constitute the final set of standard allocated line clusters. Further, the digital twin construction module 15 receives this set, first evaluates the comprehensive priority of the line features in each cluster set, and then constructs digital twin models with different accuracy levels and data update frequencies according to the priority. It should be noted that the specific process of the aforementioned sub-steps is described in S601-S608 below, and will not be repeated here.

[0072] In another possible implementation, when re-clustering all lighting lines, the line re-clustering module 14 can first calculate the local density of each lighting line in the joint feature space and the distance to higher density points to identify the cluster center. Then, the remaining points are quickly assigned to the nearest cluster center, thereby efficiently forming a set of standard assigned line clusters, which is particularly suitable for scenarios with a large number of lighting lines.

[0073] In another possible implementation, when the digital twin construction module 15 dynamically constructs a digital twin for each standard-assigned line cluster set, it can also use a simplified two-dimensional mesh diagram to express the topological relationship for low-priority cluster sets instead of constructing a detailed three-dimensional geometric model, and trigger the twin data update only when its electrical state changes beyond a threshold, thereby further reducing system resource consumption.

[0074] Therefore, the route re-clustering module 14 and the digital twin construction module 15 jointly use more comprehensive features to perform secondary clustering, resulting in more optimized route groups that better meet management needs; and implement differentiated digital twin construction strategies based on the importance of the groups, ultimately achieving economic optimization of the overall construction and operation costs of the digital twin while ensuring refined management of key areas.

[0075] Based on the above technical solution, this invention acquires current data of lighting lines and devices within a power distribution area, and sequentially performs initial clustering based on device distribution characteristics, scene feature quantification based on intra-cluster connectivity, feature correction based on inter-cluster connectivity, and secondary clustering based on dual features. Finally, it constructs a digital twin based on the importance of the clustering results. This method creatively combines the physical distribution, topological connections, and operational characteristics of the lines, dynamically and intelligently determining the appropriate granularity and precision of the digital twin's construction. This effectively optimizes the overall construction cost and resource utilization efficiency of the digital twin system while meeting the needs of power distribution automation for equipment status monitoring and management.

[0076] For example, in another embodiment of the present invention, a method for dynamically constructing digital twins of electrical equipment applicable to power distribution automation is provided. The method involves determining the distribution characteristic indicators of lighting devices contained in different lighting lines within a target area based on current data, and then clustering all lighting lines into multiple line clusters to be assigned based on these distribution characteristic indicators. Specifically, this includes the following steps: S301. Circuits containing at least one intelligent lighting device are classified as intelligent lighting circuits, and circuits without intelligent lighting devices are classified as conventional lighting circuits. Intelligent lighting devices are those capable of detecting their own current data, while conventional lighting devices are those not capable of detecting their own current data.

[0077] In this step, the line clustering module 12 analyzes the current data provided by the data acquisition module 11. Specifically, the line clustering module 12 analyzes the source characteristics of the current data: if the device current data of a lighting device is directly reported by its own sensor, then the device is determined to be an intelligent lighting device; if a lighting device does not have independent device current data, or its power consumption status can only be indirectly inferred from the current changes of its associated line, then the device is determined to be a conventional lighting device.

[0078] After identifying all lighting devices, the line clustering module 12 traverses each lighting line and counts the number of smart lighting devices it contains. If a line has 1 or more smart lighting devices, it is classified as a smart lighting line; if a line has 0 smart lighting devices, it is classified as a regular lighting line.

[0079] S302. Based on the first number of intelligent lighting devices contained in each intelligent lighting line, the total number of intelligent lighting devices contained in all intelligent lighting lines, and the positional distance between intelligent lighting devices in the intelligent lighting lines, determine the first distribution characteristic index of each intelligent lighting line.

[0080] For example, when the line clustering module 12 performs this step, it specifically includes the following steps: (1) Determine the initial scene feature value based on the first quantity, the total quantity and the location distance; wherein, the initial scene feature value is used to characterize the density and distribution characteristics of the intelligent lighting devices contained in the intelligent lighting circuit; Specifically, for any smart lighting circuit, the circuit clustering module 12 performs the following calculations: Assume the initial number of smart lighting devices in the smart lighting circuit is W, the total number of smart lighting devices in all smart lighting circuits is W1, and the distance between the w-th and w+1-th smart lighting devices in the circuit is . Where w ranges from 1 to W−1. The route clustering module 12 calculates the initial scene feature value of the route using the following formula. : Where W represents the number of smart lighting devices contained in the current smart lighting circuit (i.e., the first number). This indicates the total number of smart lighting devices contained in all smart lighting circuits; This represents the electrical or physical distance between the w-th smart lighting device and the (w+1)-th smart lighting device in the current smart lighting circuit, within the power distribution network topology. This represents the sum of distances between all adjacent smart lighting devices on the line.

[0081] It should be noted that the above formula quantifies the degree of clustering of smart lighting devices in the circuit. A larger numerator (W) indicates more smart devices in the circuit; a smaller denominator indicates a denser distribution of these devices. Both factors work together to... The larger the value of W1, the higher the density and more concentrated the distribution of the intelligent lighting devices contained in the intelligent lighting circuit, and the more obvious its scene characteristics. W1 is introduced to make relative comparisons from a global perspective.

[0082] (2) Normalize the initial scene feature values ​​to obtain the first distribution feature index.

[0083] Furthermore, the line clustering module 12 uses a maximum-minimum normalization function to calculate the initial scene feature values ​​of all intelligent lighting lines. Mapping to the interval [0, 1] yields the first distribution characteristic index q.

[0084] S303. The second number of conventional lighting devices contained in each conventional lighting circuit shall be used as the second distribution characteristic index of the conventional lighting circuit.

[0085] Specifically, for each classified conventional lighting line, the line clustering module 12 directly counts the total number of conventional lighting devices connected to that line that do not have current self-detection capabilities, and uses this number directly as the second distribution characteristic index of that line. For example, if a road lighting line connects 20 conventional streetlights, then the second distribution characteristic index of that line is 20.

[0086] S304. Based on the differences between the first distribution characteristic indicators of each intelligent lighting line, use a preset clustering algorithm to perform the first clustering of all intelligent lighting lines to form at least one first-class line cluster to be assigned.

[0087] In this step, the line clustering module 12 uses the first distribution characteristic index q of each smart lighting line as the clustering feature. The line clustering module 12 calculates the feature difference value between any two smart lighting lines i and j. (in and Let represent the first distribution characteristic indices of the i-th and j-th smart lighting lines, respectively, to obtain a difference matrix. Subsequently, the line clustering module 12 uses the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm as the preset clustering algorithm for the first clustering. The algorithm requires two preset key parameters: neighborhood radius and minimum number of samples.

[0088] For example, based on the distribution of a large amount of line feature data, the neighborhood radius can be set to 0.15, and the minimum sample size to 2. The selection rule is as follows: by analyzing the numerical distribution density of the first distribution feature index q in historical data, a neighborhood radius value is selected that allows most lines with similar q values ​​to be grouped into the same cluster, while effectively distinguishing lines with significant differences. The minimum sample size is set to 2 to ensure that an effective cluster contains at least two lines, avoiding the generation of too many clusters containing only a single line. The DBSCAN algorithm, based on density reachability, groups smart lighting lines with similar q values ​​into the same cluster, and each resulting cluster is a first-class cluster of lines to be assigned. Noise points (lines that cannot be assigned to any cluster) can be processed separately or assigned to a specific cluster.

[0089] S305. Based on the differences between the second distribution characteristic indicators of each conventional lighting line, a preset clustering algorithm is used to perform a second clustering of all conventional lighting lines to form at least one second-class line cluster to be assigned.

[0090] In this step, the line clustering module 12 uses the second distribution characteristic index (i.e., the number of conventional lighting devices) of each conventional lighting line as the clustering feature. The line clustering module 12 calculates the feature difference value between any two conventional lighting lines m and n. Where N represents the number of devices, and Let m and n represent the second distribution characteristic indicators (i.e., the number of conventional lighting devices) of the m-th and n-th conventional lighting lines, respectively, to obtain the difference matrix. Subsequently, the line clustering module 12 also uses the DBSCAN algorithm as the preset clustering algorithm for the second clustering. Considering that the number of devices is an integer and may have a large range, the value of the neighborhood radius needs to be adapted to the order of magnitude.

[0091] For example, in this step, the neighborhood radius can be set to 5, and the minimum sample size to 3. The rules for setting these values ​​are: based on the common range of differences in the number of conventional lighting circuit devices within the target area, a reasonable difference in quantity is set as the neighborhood determination criterion; for example, circuits with a difference of no more than 5 devices are considered to have similar load bases and scenario requirements. The minimum sample size is set to 3 to improve cluster stability. After the algorithm runs, conventional lighting circuits with similar device numbers are aggregated into clusters, and each cluster is a second-type cluster of circuits to be assigned.

[0092] Based on the above technical solution, this invention accurately distinguishes between intelligent lighting circuits and conventional lighting circuits, and calculates a first distribution characteristic index (reflecting the density of intelligent devices) and a second distribution characteristic index (reflecting the scale of conventional devices) for their different characteristics. Then, based on these indicators, a density clustering algorithm is used for initial clustering. This method achieves objective and automated preliminary grouping of massive lighting circuits based on their inherent composition and distribution characteristics, grouping circuits with similar functions and scene characteristics into the same cluster to be assigned. This lays a reasonable structural foundation for subsequent more refined scene feature analysis and digital twin construction, effectively avoiding the unreasonable granularity problem caused by a one-size-fits-all approach.

[0093] For example, in another embodiment of the present invention, a method for dynamically constructing a digital twin of electrical equipment applicable to power distribution automation is provided. Based on the connection relationship between lighting lines within each cluster of lines to be allocated, the power scene characteristic indicators of each lighting line are determined. Specifically, this includes the following steps: S401. Obtain the lighting devices located before the overlapping node in the current lighting line, and form a sequence of overlapping lighting devices. Wherein, the overlapping node is the physical connection point between the current lighting line and the overlapping lighting line, and the overlapping lighting line is the lighting line that has a physical connection with the current lighting line.

[0094] Optionally, this step is performed by the internal feature calculation submodule 131 in the scene feature determination module 13.

[0095] In this step, the internal feature calculation submodule 131 analyzes any selected lighting line (hereinafter referred to as the current lighting line) within any cluster of lines to be assigned, which is divided by the line clustering module 12. First, the internal feature calculation submodule 131 queries the power distribution network topology database for all other lighting lines that have a direct physical connection with the current lighting line (e.g., connected through a switch, connector, or T-contact), and defines these lines as overlapping lighting lines, and the physical connection points as overlapping nodes.

[0096] Subsequently, for each overlapping node, the internal feature calculation submodule 131 uses that node as the dividing point to obtain all lighting devices on the current lighting line that are located upstream (power supply side) of that node, and arranges them according to their electrical order on the current lighting line to form a sequence of overlapping lighting devices for that node.

[0097] S402. Based on the number of lighting devices in the reclosing lighting device sequence and the amount of current reduction after the current lighting line passes through each reclosing node, determine the line reuse impact factor of the reclosing lighting line. The line reuse impact factor characterizes the degree of influence of the reclosing lighting line on the power distribution of the current lighting line.

[0098] For example, the internal feature calculation submodule 131 determines the line reuse impact factor of overlapping lighting lines, specifically including the following steps: (1) Calculate the initial reuse impact value based on the number of lighting devices in the overlapping lighting device sequence and the amount of current reduced after the current lighting line passes through each overlapping node.

[0099] First, the internal feature calculation submodule 131 performs the following calculations for a concurrent lighting circuit: Assuming the number of lighting devices in the sequence of overlapping lighting devices corresponding to this overlapping lighting circuit is L, and the current reduction after the current circuit passes through this node is... (Unit: Ampere) The determination method is as follows: On the current lighting circuit, taking the overlapping node as the boundary, obtain the current value of the nearest monitoring point upstream of the node (or the average value of multiple upstream monitoring points), and the current value of the nearest monitoring point downstream of the node (or the average value of multiple downstream monitoring points). This is the difference between two current values. This difference represents the total current shunted from the current lighting circuit to the reclosing lighting circuit and other downstream branches. In an equivalent implementation, if a current monitoring device is installed at the beginning of the reclosing lighting circuit, then... The monitored value can be directly taken. Under ideal conditions, this monitored value is equal to the aforementioned current difference. The internal characteristic calculation submodule 131 calculates the initial multiplexing influence value of the overlapping lighting circuit using the following formula. : Where L represents the number of lighting devices contained in the sequence of overlapping lighting devices. This represents the decrease in current value of the current circuit after passing through the nth overlapping node. This value characterizes the energy loss caused by the current diversion at that node.

[0100] It should be noted that the above formula quantifies the intensity of the "demand" for power distribution from upstream lines by the overlapping lighting lines. The larger L is, the longer the line segment reused (shared) by the overlapping line and the more devices it serves. This indicates the amount of current reduction in the current flowing through the current lighting circuit after passing through the overlapping node, i.e., the amount of current diverted to the overlapping lighting circuit. The larger the value, the more electrical energy the overlapping line draws from the upstream line, and the stronger its influence on the upstream power distribution. Multiplying these two parts gives... The larger the value, the greater the influence of the overlapping lighting circuit on the power distribution of the current lighting circuit.

[0101] (2) Normalize the initial reuse impact value to obtain the line reuse impact factor.

[0102] Furthermore, the internal feature calculation submodule 131 uses a maximum-minimum normalization function to calculate the initial multiplexing influence values ​​of all overlapping lighting circuits. Mapping to the [0, 1] interval yields the normalized line reuse impact factor β.

[0103] S403. Based on the mean current change transmitted from the sequence of overlapping lighting devices to the overlapping lighting line, the line reuse influence factor, the mean first variance of the current of the sequence of overlapping lighting devices, the mean second variance of the current of the overlapping lighting line, and the positional distance between the lighting devices on both sides of the overlapping node, determine the power scene characteristic indicators of the current lighting line.

[0104] Furthermore, after obtaining the line reuse influence factor β, the internal feature calculation submodule 131, in conjunction with other electrical and spatial parameters, performs the following calculations to first determine the power scene characteristic index E of the overlapping lighting lines: Assuming that the average change in current is transmitted when the sequence of reclosing lighting devices is transmitted to the reclosing lighting circuit, the mean value is... (Unit: Amperes); The mean variance of the current data for the sequence of overlapping lighting devices is _____. The first variance mean is calculated as follows: for each lighting device in the overlapping lighting device sequence, current data is collected at each sampling time within a preset time period. The variance of the current data for each lighting device is calculated, and then the arithmetic mean of these variances is obtained. The variance mean of the current data of the overlapping lighting circuit itself is... The mean of the second variance is calculated as follows: For each lighting device on the overlapping lighting line, current data is collected at each sampling time within the same preset time period. The variance of the current data for each lighting device is calculated, and then the arithmetic mean of these variances is obtained. The positional distance between the lighting devices on both sides of the overlapping node (i.e., the last device in the overlapping lighting device sequence and the first device in the overlapping lighting line) is... (Unit: meters). The internal feature calculation submodule 131 calculates the power scene characteristic index E of the overlapping lighting circuits using the following formula: Where E represents the power consumption characteristic index of overlapping lighting circuits. β represents the long-term average value of the change in current as it flows from the sequence of repeating lighting devices into the repeating lighting circuit. This value reflects the average intensity of energy transfer. β represents the calculated circuit reuse influence factor. The larger the value, the stronger the influence.

[0105] It should be noted that the above formula consists of two parts. The first part... The physical meaning of is to measure the net change in electrical energy transferred from the multiplexing section to the line after eliminating the contribution of the line reuse effect (β). Part Two The physical meaning is to compare the current fluctuation difference between the multiplex section and the line itself. ), and combined with the absolute fluctuation level of the line itself ( The denominator is denoted by distance d1, which represents the uniqueness of the power demand of the line. The distance d1 in the denominator is to account for the potential attenuation effect of electrical distance on wave propagation; the greater the distance, the more independent the line's characteristics may be. The E value obtained by multiplying the two parts comprehensively reflects the unique power demand and wave characteristics exhibited by this overlapping lighting line under the influence and constraints of the upstream multiplexing section. The larger this value, the more distinct and independent the line's scene characteristics.

[0106] Furthermore, after determining the power scene characteristic index E of the overlapping lighting lines, the scene feature determination module 13 synthesizes the power scene characteristic index E calculated from all directly connected overlapping lighting lines (for example, by taking the arithmetic mean or weighted average), and determines the synthesized result as the power scene characteristic index of the current lighting line. Then, the maximum-minimum value normalization function is used to map the power scene characteristic index of the current lighting line to the interval [0, 1] to obtain the normalized power scene characteristic index e of the current lighting line.

[0107] Based on the above technical solution, this invention accurately analyzes the topological connections between lighting lines, quantifies the impact factors of line reuse, and comprehensively considers multi-dimensional parameters such as current variation, fluctuation characteristics, and spatial distance to ultimately determine the power scenario characteristic indicators of each line. This method achieves a deep mapping from physical connections to functional characteristics, transforming complex power grid topology and operational data into quantitative indicators characterizing the personalized scenario requirements of each line. This provides an accurate and reliable data foundation for subsequent line re-clustering based on scenario similarity and the construction of differentiated digital twins.

[0108] For example, in another embodiment of the present invention, a method for dynamically constructing a digital twin of electrical equipment applicable to power distribution automation is provided. Based on the connection relationship of lighting circuits between different clusters of circuits to be allocated, the characteristic indicators of the power scene are corrected to obtain corrected power scene characteristic indicators. Specifically, this includes the following steps: S501. Among all the circuit clusters to be assigned, identify other lighting circuits that are directly connected to the current lighting circuit as cross-cluster connection circuits.

[0109] Optionally, this step is performed by the cross-cluster correction submodule 132 in the scene feature determination module 13.

[0110] In this step, the cross-cluster correction submodule 132 takes the current lighting line, whose power scene characteristic indicators have been determined in step S203, as the analysis object. The cross-cluster correction submodule 132 first obtains the complete set of all unassigned line clusters generated by the line clustering module 12. This complete set includes the first and second types of unassigned line clusters mentioned above. Then, the cross-cluster correction submodule 132 traverses this set, and for the current lighting line, starting from its own unassigned line cluster, sequentially retrieves the lighting line list within each of the remaining unassigned line clusters.

[0111] Understandably, the cross-cluster correction submodule 132 queries the distribution network topology database, matching the unique identifier of the current lighting line with line identifiers in these lists to check for any documented direct physical connections between them (e.g., via shared switch cabinets, connectors, or T-junctions). All lighting lines identified as having a direct physical connection to the current lighting line and located in other unassigned line clusters are marked as cross-cluster connected lines. This step ensures that all physical connections beyond the initial cluster boundaries are discovered.

[0112] S502. In the cross-cluster connection lines, the lines that do not belong to the current lighting line to be assigned are taken as reference lighting lines, and other to-be-assigned line clusters distributed by the reference lighting lines that are different from the current lighting line to be assigned are taken as reference clusters.

[0113] Furthermore, the cross-cluster correction submodule 132 filters the list of cross-cluster connection lines identified in S501. Since the definition of "cross-cluster connection line" implicitly means that it does not belong to the cluster to which the current lighting line belongs, this step is essentially a confirmation and statistical analysis. The cross-cluster correction submodule 132 formally identifies each line in the list as a reference lighting line. At the same time, the cross-cluster correction submodule 132 extracts the cluster identifier of the line to be assigned to each reference lighting line.

[0114] To avoid duplication, the cross-cluster correction submodule 132 performs deduplication on these cluster identifiers, ultimately obtaining a unique set of other line clusters to be assigned that are different from the current cluster, and defines this set of clusters as reference clusters. For example, if the current lighting line belongs to cluster A, and its cross-cluster connecting lines belong to clusters B and C respectively, then the reference lighting lines are these lines from clusters B and C, and the reference cluster is {B, C}.

[0115] S503. Based on the number of reference lighting lines and reference clusters, the power scene characteristic indicators of the current lighting line are numerically corrected to obtain the corrected power scene characteristic indicators.

[0116] For example, the cross-cluster correction submodule 132 corrects the power scene characteristic indicators of the current lighting circuit using the following formula: Assume the normalized power scene characteristic index of the current lighting circuit determined by S203 is e, the number of reference lighting circuits obtained in S502 is M, and the number of reference clusters is M1. The cross-cluster correction submodule 132 first calculates the sum of M and M1, S = M + M1. Then, the cross-cluster correction submodule 132 normalizes this sum S to obtain norm(S). Finally, the corrected power scene characteristic index e1 is calculated using the following formula: Understandably, the above formula quantifies the additional impact on the characteristics of the lighting scene due to cross-cluster connections. M and M1 measure the complexity of cross-cluster connections from two dimensions: the number of connected lines and the breadth of affected clusters, respectively. The larger the sum of M and M1, the more extensive and complex the coupling between the current lighting line and external clusters. The normalization term norm(S) reflects the general level of this cross-cluster coupling strength relative to all lines. As a multiplier applied to the original index e, this means that the more extensive and complex the cross-cluster connections (the larger the norm(S)), the more the scenario characteristics represented by the original index e need to be "amplified" or "enhanced," because the operating state of the line may be subject to more external constraints or influences, and its functional uniqueness or sensitivity in the overall network may be higher. The resulting e1 is the corrected power scenario characteristic index, which contains comprehensive information on intra-cluster characteristics and cross-cluster coupling effects.

[0117] Based on the above technical solution, this invention systematically identifies and quantifies the physical connection relationship between the current lighting line and the external cluster of lines to be allocated, and accordingly weights and corrects the initial indicators characterizing its scene features. This method effectively compensates for the limitations of analysis based solely on intra-cluster analysis, enabling the feature indicators used for line re-clustering to more comprehensively and accurately reflect the actual coupling environment of each lighting line in the overall power distribution network topology and the degree of external influence, thereby providing key data augmentation for forming a more reasonable and stable set of standard allocation line clusters.

[0118] For example, in another embodiment of the present invention, a method for dynamically constructing digital twins of electrical equipment applicable to power distribution automation is provided. This method involves re-clustering all lighting lines based on distribution characteristic indicators and modified power scenario characteristic indicators to form multiple standard allocation line cluster sets. A digital twin is then dynamically constructed for each standard allocation line cluster set. Specifically, this includes the following steps: S601. Determine the first comprehensive difference degree based on the difference degree of the first distribution characteristic index and the difference degree of the corrected power scene characteristic index between each pair of intelligent lighting circuits.

[0119] In this step, the line re-clustering module 14 obtains the first distribution feature index qi and qj (i.e., the first distribution feature index of each line determined in step S302) and the corrected power scene feature index for any two smart lighting lines i and j. and The line re-clustering module 14 calculates the difference of the first distribution characteristic index. =∣qi-qj∣, calculate the difference degree of the corrected power scenario characteristic index. =∣ - | Then, the line re-clustering module 14 calculates the first comprehensive difference degree. , which serves as the mean of the difference between the two.

[0120] Understandable, the first comprehensive degree of difference By taking the arithmetic mean of the two dissimilarity values, a comprehensive measure of the dissimilarity between a pair of smart lighting lines in terms of "device distribution" and "scene characteristics" is obtained. The larger the value, the less similar the two lines are, and the more likely they are to belong to different clusters in clustering.

[0121] S602. Based on the first comprehensive difference degree, use a preset clustering algorithm to perform a third clustering on all smart lighting circuits to form at least one smart standard cluster.

[0122] Furthermore, the line re-clustering module 14 uses all smart lighting lines as sample points, and calculates the first comprehensive difference between each pair of sample points. As a distance metric, a distance matrix is ​​constructed. The line re-clustering module 14 performs a third clustering using a preset clustering algorithm. For example, the preset clustering algorithm is the DBSCAN algorithm. The DBSCAN algorithm requires two preset parameters: neighborhood radius and minimum number of samples.

[0123] For clustering smart lighting lines, for example, based on the distribution characteristics of the first comprehensive dissimilarity in historical data, a neighborhood radius of 0.2 and a minimum sample size of 3 are set. The selection rule is: by analyzing the statistical distribution (e.g., quantiles) of all first comprehensive dissimilarity values, a value is selected that allows most similar lines ( Small clusters are grouped into the same cluster, while effectively separating different clusters. The neighborhood radius value of j (largest); the minimum number of samples is set to 3 to ensure that each cluster contains at least 3 lines, avoiding the generation of clusters that are too small or noise points. The line re-clustering module 14 runs the DBSCAN algorithm to divide the smart lighting lines into several clusters, each of which is called a smart standard cluster.

[0124] S603. Determine the second comprehensive difference degree based on the difference degree of the second distribution characteristic index between each pair of conventional lighting circuits and the difference degree of the modified power scene characteristic index.

[0125] Specifically, for any two conventional lighting lines m and n, the line re-clustering module 14 obtains their second distribution characteristic index (i.e., the number of conventional lighting devices). and And correct the characteristic indicators of power scenarios and The line re-clustering module 14 calculates the dissimilarity of the second distribution characteristic index. =∣ - | Calculate the difference in characteristic indicators of the corrected power scenario =∣ - | Then, the line re-clustering module 14 calculates the second comprehensive dissimilarity. , which serves as the mean of the difference between the two.

[0126] Understandably, the second comprehensive degree of difference By taking the arithmetic mean of the two dissimilarity values, a comprehensive measure of the difference between a pair of conventional lighting lines in terms of "installation size" and "scene characteristics" is obtained. The larger the value, the less similar the two lines are.

[0127] S604. Based on the second comprehensive difference degree, use the preset clustering algorithm to perform a fourth clustering on all conventional lighting circuits to form at least one conventional standard cluster.

[0128] Furthermore, the line re-clustering module 14 uses all conventional lighting lines as sample points, and uses the second comprehensive difference between each pair of sample points. As a distance metric, a distance matrix is ​​constructed. The line re-clustering module 14 uses a preset clustering algorithm for the fourth clustering. For example, the preset clustering algorithm also adopts the DBSCAN algorithm. For the clustering of regular lighting lines, the parameters are set to a neighborhood radius of 5 and a minimum number of samples of 4. The rules for their values ​​are: considering that the number of regular lighting devices is an integer and may vary greatly, a neighborhood radius of 5 means that lines with a difference of 5 in the number of devices are allowed to be considered neighbors in the density calculation; the minimum number of samples is set to 4, which is set slightly higher to improve the stability of the clusters and avoid too many small clusters due to fluctuations in the number of lines. The line re-clustering module 14 runs the DBSCAN algorithm to divide the regular lighting lines into several clusters, each of which is called a regular standard cluster.

[0129] S605. Based on the intelligent standard cluster and the conventional standard cluster, determine a set of multiple standard allocation line clusters.

[0130] Finally, the line re-clustering module 14 merges all the smart standard clusters formed in S602 and all the regular standard clusters formed in S604. Each smart standard cluster or regular standard cluster is itself a set of lighting lines. The line re-clustering module 14 defines the entirety of these clusters as a set of multiple standard allocation line clusters. In other words, the standard allocation line cluster set = {all smart standard clusters} ∪ {all regular standard clusters}. This set serves as the final grouping result and is used for the subsequent construction of the digital twin.

[0131] S606. Based on the statistical characteristics of the distribution characteristics of lighting lines and the modified power scenario characteristics within each standard-allocated line cluster set, determine the comprehensive priority of each standard-allocated line cluster set.

[0132] In this step, the digital twin construction module 15 assigns a standard line cluster set to each standard line cluster. Perform the following operations to determine its overall priority. : (1) Get the set Distribution characteristics of all lighting circuits and corrected power scenario characteristics e1.

[0133] (2) Standardize the dimensions of the distribution characteristic indicators: For intelligent lighting circuits, the first distribution characteristic indicator q is already in the [0, 1] interval and can be used directly; for conventional lighting circuits, the second distribution characteristic indicator N (number of devices) needs to be normalized. The normalization method is the same as the normalization method introduced above, and will not be repeated here. After this processing, the distribution characteristic indicator DF of all circuits falls in the [0, 1] interval.

[0134] (3) Calculate the set Arithmetic mean of DF for all lines within the network .

[0135] (4) Calculate the set The arithmetic mean of the corrected power scenario characteristic index e1 for all lines within the system. .

[0136] (5) Calculate the overall priority using the following weighted formula. : The values ​​of 0.4 and 0.6 are exemplary and can be adjusted in practical applications based on the emphasis placed on device distribution characteristics and scene characteristics. The calculated values ​​are... The value range is also [0, 1], and the larger the value, the higher the overall priority of the set.

[0137] S607. For a set of standard-assigned line clusters with a comprehensive priority greater than a preset priority threshold, construct a digital twin with first precision and first update frequency.

[0138] Furthermore, the digital twin construction module 15 sets a preset priority threshold T, exemplarily set to T=0.5. The rule for its selection is: based on a large amount of historical data or expert experience, a boundary value is selected that can roughly divide the standard allocation line cluster set into two categories: "high importance" and "general importance." Typically, the median of the overall priority of all sets or a certain empirical percentile (such as the 60th percentile) can be selected as the threshold. For those meeting the requirements... For each standard assigned circuit cluster set of T, the digital twin construction module 15 calls high-precision modeling tools (such as 3D reconstruction software based on laser point clouds or oblique photogrammetry) to construct a digital twin for it. The geometric model is required to have a Level of Detail (LOD) of 300 or higher, containing detailed component geometry and texture information. Simultaneously, a real-time data channel is configured to synchronously update the twin's operational status data (such as current, voltage, and switching status) at a first update frequency (e.g., once per second).

[0139] S608. For a set of standard assigned line clusters with a comprehensive priority less than or equal to a preset priority threshold, construct a digital twin with a second precision and a second update frequency. The first precision is higher than the second precision, and the first update frequency is higher than the second update frequency.

[0140] Specifically, the digital twin building block 15 is for meeting the following requirements: For each standard-assigned line cluster set ≤T, the digital twin construction module 15 calls a lightweight modeling tool (such as a simplified two-dimensional topology diagram or a low-polygon three-dimensional model) to construct a digital twin for it. The model accuracy requirement is second precision, for example, a LOD level of 100, which only expresses the basic outline and topological connection relationship of the equipment. Data updates adopt a second update frequency, such as once per minute, or are only triggered when a change event occurs in the equipment status (such as a switch change).

[0141] Based on the above technical solution, this embodiment of the invention performs secondary fine clustering by integrating the device distribution characteristics of the lines with the corrected scene characteristics, forming a more reasonable set of standard-allocated line clusters; and configures the accuracy and update frequency of the digital twin differently according to the comprehensive priority of the line characteristics within the set. This method ensures refined and real-time management of high-priority areas while significantly reducing the modeling and maintenance costs of low-priority areas, achieving optimized allocation of digital twin construction resources in the power distribution automation system.

[0142] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0143] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A method for dynamically constructing digital twins of electrical equipment suitable for power distribution automation, characterized in that, The method includes: Acquire current data within the target area; wherein, the current data includes line current data of each lighting circuit within the target area and device current data of the corresponding lighting device; Based on the current data, the distribution characteristic index of the lighting devices contained in different lighting lines within the target area is determined, and all lighting lines are clustered into multiple line clusters to be assigned based on the distribution characteristic index. Based on the connection relationship between each lighting line within each cluster of lines to be allocated, determine the power scene characteristic indicators of each lighting line; Based on the connection relationship of lighting circuits between different clusters of circuits to be allocated, the power scene characteristic index is corrected to obtain the corrected power scene characteristic index. Based on the distribution characteristic index and the modified power scene characteristic index, all lighting lines are re-clustered to form multiple standard allocation line cluster sets, and a digital twin is dynamically constructed for each standard allocation line cluster set.

2. The method for dynamically constructing digital twins of electrical equipment suitable for power distribution automation according to claim 1, characterized in that, The distribution characteristic indicators include a first distribution characteristic indicator and a second distribution characteristic indicator; the distribution characteristic indicators of lighting devices contained in different lighting circuits within the target area are determined based on the current data, specifically including: Each lighting circuit containing at least one intelligent lighting device is classified as an intelligent lighting circuit, and each circuit without an intelligent lighting device is classified as a conventional lighting circuit. Among them, an intelligent lighting device is a lighting device that has the ability to detect its own current data, and a conventional lighting device is a lighting device that does not have the ability to detect its own current data. The first distribution characteristic index of each intelligent lighting line is determined based on the first number of intelligent lighting devices contained in each intelligent lighting line, the total number of intelligent lighting devices contained in all intelligent lighting lines, and the positional distance between intelligent lighting devices in the intelligent lighting lines. The second number of conventional lighting devices contained in each conventional lighting line is used as the second distribution characteristic index of the conventional lighting line.

3. The method for dynamically constructing digital twins of electrical equipment suitable for power distribution automation according to claim 2, characterized in that, Determining the first distribution characteristic index for each of the intelligent lighting circuits specifically includes: An initial scene feature value is determined based on the first quantity, the total quantity, and the location distance; wherein, the initial scene feature value is used to characterize the density and distribution characteristics of the intelligent lighting devices contained in the intelligent lighting circuit; The initial scene feature values ​​are normalized to obtain the first distribution feature index.

4. The method for dynamically constructing digital twins of electrical equipment suitable for power distribution automation according to claim 2, characterized in that, The types of the multiple unassigned line clusters include a first type of unassigned line clusters and a second type of unassigned line clusters; Based on the aforementioned distribution characteristic indicators, all lighting lines are clustered into multiple line clusters to be assigned, specifically including: Based on the differences between the first distribution characteristic indicators of each smart lighting line, a preset clustering algorithm is used to perform a first clustering of all smart lighting lines to form at least one first-class line cluster to be assigned. Based on the differences between the second distribution characteristic indicators of each conventional lighting line, the preset clustering algorithm is used to perform a second clustering on all conventional lighting lines to form at least one second-class line cluster to be assigned.

5. The method for dynamically constructing digital twins of electrical equipment suitable for power distribution automation according to claim 1, characterized in that, Based on the connection relationships between the lighting lines within each of the proposed lighting clusters, the power scene characteristic indicators for each lighting line are determined, specifically including: Obtain the lighting devices located before the overlapping node in the current lighting line to form a sequence of overlapping lighting devices; wherein, the overlapping node is the physical connection point between the current lighting line and the overlapping lighting line, and the overlapping lighting line is a lighting line that has a physical connection with the current lighting line. Based on the number of lighting devices in the overlapping lighting device sequence and the amount of current reduction after the current lighting line passes through each overlapping node, the line reuse influence factor of the overlapping lighting line is determined; wherein, the line reuse influence factor is used to characterize the degree of influence of the overlapping lighting line on the power distribution of the current lighting line. The power scene characteristic index of the current lighting line is determined based on the average current change transmitted from the overlapping lighting device sequence to the overlapping lighting line, the line reuse influence factor, the average first variance of the current of the overlapping lighting device sequence, the average second variance of the current of the overlapping lighting line, and the positional distance between the lighting devices on both sides of the overlapping node.

6. The method for dynamically constructing digital twins of electrical equipment suitable for power distribution automation according to claim 5, characterized in that, Determining the line reuse impact factor of the overlapping lighting lines specifically includes: The initial reuse impact value is calculated based on the number of lighting devices in the overlapping lighting device sequence and the amount of current reduction after the current lighting line passes through each overlapping node. The initial multiplexing impact value is normalized to obtain the line multiplexing impact factor.

7. The method for dynamically constructing digital twins of electrical equipment suitable for power distribution automation according to claim 1, characterized in that, Based on the connection relationship of lighting circuits among the different clusters of circuits to be allocated, the power scene characteristic index is corrected to obtain the corrected power scene characteristic index, which specifically includes: Identify other lighting lines that are directly connected to the current lighting line from all the line clusters to be assigned as cross-cluster connection lines; In the cross-cluster connection lines, the lines that do not belong to the assigned line cluster to which the current lighting line belongs are taken as reference lighting lines, and other assigned line clusters that are distributed by the reference lighting lines but are different from the assigned line cluster to which the current lighting line belongs are taken as reference clusters. Based on the number of the reference lighting lines and the reference clusters, the power scene characteristic indicators of the current lighting lines are numerically corrected to obtain the corrected power scene characteristic indicators.

8. The method for dynamically constructing digital twins of electrical equipment suitable for power distribution automation according to claim 1, characterized in that, Based on the distribution characteristic index and the modified power scenario characteristic index, all lighting lines are re-clustered to form multiple standard allocation line clusters, specifically including: The first comprehensive difference is determined based on the difference between the first distribution characteristic index and the difference between the corrected power scene characteristic index between each pair of smart lighting lines; Based on the first comprehensive difference degree, a third clustering is performed on all smart lighting circuits using a preset clustering algorithm to form at least one smart standard cluster. The second comprehensive difference is determined based on the difference between the second distribution characteristic index and the difference between the modified power scene characteristic index between each pair of conventional lighting circuits; Based on the second comprehensive difference degree, the preset clustering algorithm is used to perform a fourth clustering on all conventional lighting circuits to form at least one conventional standard cluster. Based on the intelligent standard cluster and the conventional standard cluster, a set of multiple standard allocation line clusters is determined.

9. The method for dynamically constructing digital twins of electrical equipment suitable for power distribution automation according to claim 1, characterized in that, A digital twin is dynamically constructed by allocating a set of line clusters for each standard, specifically including: Based on the statistical characteristics of the distribution characteristic index and the modified power scene characteristic index of the lighting lines within each standard allocation line cluster set, the comprehensive priority of each standard allocation line cluster set is determined. For the standard line cluster set whose comprehensive priority is greater than a preset priority threshold, construct a digital twin with a first precision and a first update frequency; For a set of standard assigned line clusters with a comprehensive priority less than or equal to the preset priority threshold, a digital twin with a second precision and a second update frequency is constructed; wherein, the first precision is higher than the second precision, and the first update frequency is higher than the second update frequency.

10. The method for dynamically constructing digital twins of electrical equipment suitable for power distribution automation according to any one of claims 1-9, characterized in that, Acquire current data within the target area, specifically including: Based on the preset regional division rules, the target region is divided into multiple analysis sub-regions; Within each analysis sub-region, device current data of each lighting device and line current data of the corresponding lighting circuit are collected at a preset sampling frequency.

Images