Power utilization information feature optimization method and system considering power supply scene and topological relation

Abstract

The invention discloses a power consumption information feature optimization method and system considering a power supply scene and a topological relation, and relates to the technical field of power consumption information monitoring and processing of a power distribution network, and the method comprises the steps: obtaining the power consumption information data of measurement points in a power supply region, and obtaining the power consumption information of the measurement points based on the similarity and difference correlation between the operation data and configuration information between the measurement points; the method comprises the following steps: constructing a power utilization information characteristic mapping relation reflecting a power supply topology structure, determining a power supply scene type according to a hierarchical architecture of power supply topology and load space-time distribution characteristics, extracting power supply characteristics and power utilization rules, establishing a characteristic optimization model based on a power supply scene, executing a dual screening mechanism, generating a cross fusion characteristic set, and obtaining a power supply scene model; and injecting confusion variables into the fusion feature set to form an extended feature library, evaluating an optimization effect based on a confusion resistance index, and triggering iterative optimization if the optimization effect does not reach the standard. According to the method, the defect of relatively large feature dimension is overcome, the accuracy and practicability of power utilization information feature optimization are effectively improved, and efficient data support is provided for power supply monitoring operation management.

Description

Technical Field

[0001] The present invention relates to the technical field of power consumption information monitoring and processing in a distribution network, and in particular to a method and system for optimizing power consumption information characteristics by considering power supply scenarios and topological relationships. Background Art

[0002] As the terminal link of the power system, the distribution network carries a massive amount of user load data and equipment operation information. This data contains implicit user power consumption behavior, equipment operating status, and network topology characteristics, and is the source of data information supporting power supply monitoring and operation management. However, the distribution network has a complex topology, highly random power consumption behavior, a large number of measurement devices, and a large amount of data. Acquiring power consumption information characteristics is difficult, and existing methods have technical bottlenecks:

[0003] Current methods fail to account for differences in power supply and usage scenarios, such as dense urban loads and complex usage patterns, and long rural lines and dispersed loads. Power supply and usage patterns vary significantly across different scenarios, and failure to account for these changes and adaptability can negatively impact the accuracy of data analysis and processing applications.

[0004] Most existing technologies do not consider the relationship between the measurement data and the measurement points, reflecting the spatiotemporal relationships such as similarities, differences and hierarchies of power consumption behaviors under the constraints of the power supply topology. If these factors are not taken into account, it is impossible to reflect the potential constraints, associations and changing patterns of power consumption information.

[0005] With the increase in the amount of electricity consumption information collected and the improvement of load data dimensions, traditional data information feature processing methods often ignore the inherent correlation between data and the interactive nature of feature coupling. It is necessary to make full use of the potential rules and constraints of data information scenarios to form optimized data information features that reflect physical properties, which will help to improve feature value and model performance through elimination, screening, and interactive fusion. Summary of the Invention

[0006] In view of the above-mentioned existing problems, the present invention provides a method and system for optimizing electricity consumption information characteristics taking into account power supply scenarios and topological relationships, so as to solve the problems in the existing technology that the differences in power supply and electricity consumption scenarios are not taken into account, the data information analysis and processing applications are inaccurate, and the potential constraints, associations and change laws of electricity consumption information cannot be reflected.

[0007] To solve the above technical problems, a power consumption information feature optimization method considering power supply scenarios and topological relationships is proposed, including:

[0008] The power consumption information data of the measuring points within the power supply area is obtained, and based on the similarities and differences between the operating data and configuration information between the measuring points, a mapping relationship of the power consumption information characteristics reflecting the power supply topology is constructed; according to the hierarchical architecture of the power supply topology and the spatiotemporal distribution characteristics of the load, the power supply scenario type is determined, and the power supply characteristics and power consumption patterns are extracted; a feature optimization model based on the power supply scenario is established, and a double screening mechanism is implemented by dynamically adjusting the feature weights and combining physical constraints with weight threshold limits to generate a cross-fusion feature set; confusion variables are injected into the fused feature set to form an extended feature library, and the optimization effect is evaluated based on the confusion resistance index. If the standard is not met, iterative optimization is triggered.

[0009] As a preferred solution of the method for optimizing power consumption information characteristics considering power supply scenarios and topological relationships described in the present invention, wherein: the power consumption information data of the measuring points includes calculating the voltage, current and power flow between the measuring points;

[0010] The operation data correlation includes calculating the operation data correlation through a similarity measurement method between measuring points; the construction of a mapping relationship of power consumption information characteristics reflecting the power supply topology structure includes: the similarity measurement method between measuring points includes combining line impedance and load fluctuation factors with dynamic weighted analysis, and expanding the measurement point connection relationship layer by layer based on a recursive algorithm to construct a power supply topology network model reflecting the hierarchical structure and spatiotemporal distribution.

[0011] As a preferred solution of the power consumption information feature optimization method considering power supply scenarios and topological relationships described in the present invention, the dynamic weighted analysis includes linearly weighting the exponential decay term of the voltage mean difference between the measuring points with the Pearson correlation coefficient. When the weighted similarity measure exceeds a preset threshold, the measuring points are determined to be located in the same branch, and a hierarchical connection relationship is established based on the fundamental wave phase difference.

[0012] Among them, the weight of the voltage mean difference term is smaller than the weight of the correlation coefficient term.

[0013] As a preferred embodiment of the method for optimizing power consumption information characteristics considering power supply scenarios and topological relationships described in the present invention, the method includes extracting power supply characteristics and power consumption patterns, determining the scenario type based on the depth, branch node density, and global density of the power supply topology network model. When the network depth reaches a preset level and the branch node density exceeds a threshold, the scenario is determined to be a load-intensive scenario; otherwise, the scenario is determined to be a load-distributed scenario.

[0014] Among them, the branch node density is calculated by the ratio of the number of branch nodes in each layer to the total number of nodes, and the global density is calculated by the ratio of the number of branch nodes in the entire network to the total number of nodes.

[0015] As a preferred solution of the power consumption information feature optimization method considering power supply scenarios and topological relationships described in the present invention, wherein: the establishment of a feature optimization model based on the power supply scenario includes dividing the power consumption information features into two categories: character type and numerical type, evaluating the feature importance through chi-square test and normalized covariance correlation, and dynamically updating the feature weights in combination with the scenario type adjustment coefficient, setting a convergence condition to control the number of iterations until the difference in weights of adjacent iterations is lower than a preset threshold;

[0016] Among them, the normalized covariance correlation includes calculating the covariance of the numerical features of the measurement points and the branch node density of the hierarchy, and generating the final weight in combination with the feature importance score.

[0017] As a preferred solution of the method for optimizing electricity consumption information features considering power supply scenarios and topological relationships described in the present invention, the execution of a double screening mechanism includes setting the first screening condition as the feature weight being lower than a preset threshold, and the second screening condition as the feature data violating physical constraints and the error exceeding the tolerance range, retaining features that meet both the first and second screening conditions, and eliminating redundant and edge features.

[0018] As a preferred solution of the power consumption information feature optimization method considering power supply scenarios and topological relationships described in the present invention, wherein: forming an extended feature library includes proportionally injecting irrelevant environmental variables as confounding variables into the cross-fusion feature set to generate an extended feature library, evaluating the retention of the confounding variables through feature weight calculation, and quantifying the optimization effect based on the confusion resistance index formula;

[0019] The formula for the confusion resistance index is:

[0020]

[0021] Among them, CRI is the confounding resistance index, RSC is the number of retained cross-fusion features, OSC is the number of original cross-fusion features, CE is the number of confounding variables correctly eliminated, CS is the total number of injected confounding variables, and SF is the total number of features. When CRI>1.5, it indicates that the optimization effect is good. When CRI≤1.5, the feature optimization model is triggered and it continues to iterate, readjust the model and improve the model parameters.

[0022] The iterative optimization includes adjusting the complexity penalty coefficient and the scene type adjustment coefficient of the feature weight model according to the confusion resistance index result, and re-executing the feature weight calculation and double screening until the confusion resistance index reaches a preset target value.

[0023] As a preferred solution of the electricity consumption information feature optimization system considering power supply scenarios and topological relationships described in the present invention, it is characterized by including a data acquisition and preprocessing module, a power supply scenario identification and feature extraction module, a feature optimization model construction and screening module, and an extended feature library generation and optimization effect evaluation module.

[0024] The data acquisition and preprocessing module includes a measuring point electricity consumption information acquisition unit and an operation data correlation analysis unit, which are used to obtain electricity consumption information data of each measuring point in the power supply area, calculate the operation data correlation between the measuring points through a similarity measurement method, and construct a power consumption information characteristic mapping relationship reflecting the power supply topology structure.

[0025] The power supply scenario identification and feature extraction module includes a power supply topology analysis unit, a scenario type determination unit, and a power supply characteristic and power consumption pattern extraction unit. It is used to determine the power supply scenario type based on the hierarchical architecture of the power supply topology and the temporal and spatial distribution characteristics of the load, extract the power supply characteristics and power consumption patterns related to the power supply scenario, and assist in scenario type determination through network depth, branch node density and global density indicators.

[0026] The feature optimization model construction and screening module includes a feature classification unit, a feature importance evaluation unit, a feature weight dynamic adjustment unit, and a dual screening mechanism execution unit, which is used to divide the electricity consumption information features into two categories: character type and numerical type, and evaluate the feature importance through chi-square test and normalized covariance correlation respectively, dynamically update the feature weight according to the scenario type adjustment coefficient, set the convergence condition to control the number of iterations, and execute the dual screening mechanism.

[0027] The extended feature library generation and optimization effect evaluation module includes a confusion variable injection unit, an extended feature library generation unit, a confusion resistance index calculation unit, and an iterative optimization triggering unit. It is used to inject irrelevant environmental variables as confusion variables into the cross-fusion feature set to generate an extended feature library, evaluate the retention of confounding variables through feature weight calculation, and quantify the optimization effect based on the confusion resistance index formula.

[0028] A computer device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the method for optimizing power consumption information characteristics considering power supply scenarios and topological relationships are implemented.

[0029] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the method for optimizing power consumption information characteristics taking into account power supply scenarios and topological relationships.

[0030] Beneficial effects of the present invention: The present invention obtains the electricity consumption information of the required measuring points in the power supply area, and constructs a mapping relationship of electricity consumption information characteristics based on the power supply topology based on the similarities and differences of the operating data and configuration information at different measuring points. Compared with the traditional chain and star models, the present invention accurately represents the branches and hierarchical connection characteristics that are prevalent in power supply scenarios; based on the hierarchical distribution characteristics, the concentrated power supply scenario and the sparse power supply scenario are distinguished, which solves the problem that the traditional method ignores the topological relationship and scenario differences; the feature weight is dynamically adjusted through the scenario adjustment coefficient, so that the feature selection is more in line with the actual power consumption scenario, which significantly improves the model's adaptability to complex power grid environments. The double screening mechanism ensures that redundant features are eliminated and the cross-fusion features of power supply scenarios and topological relationships are retained to improve the accuracy of data information analysis and processing applications, and provide better data support for power supply operation and management services. Correspondingly, the present invention also proposes a confusion resistance index evaluation index that can be used to evaluate the optimization method, verify the reliability of the optimization method and improve the feature weight model according to the confusion resistance index, and construct a double screening mechanism for features that considers topological relationships and power consumption scenarios, which solves the defect of large feature dimensions, effectively improves the accuracy and practicality of power consumption information feature optimization, and provides efficient data support for power supply monitoring and operation management. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0032] Figure 1 This is a general flow chart of a method for optimizing power consumption information characteristics taking into account power supply scenarios and topological relationships, provided by one embodiment of the present invention.

[0033] Figure 2 A flow chart of a system solution for an electricity usage information feature optimization system that considers power supply scenarios and topology relationships, provided in accordance with an embodiment of the present invention. DETAILED DESCRIPTION

[0034] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0035] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0036] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it individually or selectively refer to an embodiment that is mutually exclusive of other embodiments.

[0037] The present invention is described in detail with reference to schematic diagrams. For ease of illustration, cross-sectional views of device structures may be partially enlarged and not to scale when describing embodiments of the present invention. Furthermore, the schematic diagrams are merely illustrative and should not limit the scope of the present invention. Furthermore, in actual production, the three-dimensional dimensions of length, width, and depth should be included.

[0038] In the description of the present invention, it should be noted that the terms "upper, lower, inner, and outer" and other references to orientations or positional relationships are based on the orientations or positional relationships shown in the accompanying drawings and are intended solely to facilitate and simplify the description of the present invention. They are not intended to indicate or imply that the devices or components referred to must have, be constructed, or operate in a specific orientation, and therefore should not be construed as limitations on the present invention. Furthermore, the terms "first, second, or third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0039] In this disclosure, unless otherwise specified or limited, the terms "mounted," "connected," and "connected" should be interpreted broadly. For example, they may refer to fixed, removable, or integral connections. They may also refer to mechanical, electrical, or direct connections, indirect connections through an intermediary, or internal communication between two components. Those skilled in the art will understand the specific meanings of these terms in this disclosure.

[0040] Example 1, reference Figure 1 , which is the first embodiment of the present invention, provides a method for optimizing power consumption information characteristics considering power supply scenarios and topological relationships, including:

[0041] S1: Obtain electricity consumption information data of measurement points within the power supply area, and construct a mapping relationship of electricity consumption characteristics reflecting the power supply topology based on the similarities and differences between the operating data and configuration information between the measurement points.

[0042] Based on the power consumption information of each power supply measuring point, the Pearson correlation coefficient of the voltage, current, and power flow between each measuring point is calculated. Taking into account the influence of factors such as line impedance and load fluctuation, the dynamic weighted coefficient method is used to express the similarity or difference measurement of the operating data between two measuring points, which helps to reflect the correlation characteristics of the power consumption information between the corresponding measuring points of the power supply topology. Based on the recursive algorithm, the measuring points connected to the node are searched in sequence, and then expanded layer by layer to all measuring points to construct a hierarchical structure and spatiotemporal distribution relationship based on the power supply topology.

[0043] It should be noted that the Pearson correlation coefficient between the power supply measurement point u and the power supply measurement point v is calculated based on the power consumption information of each power supply measurement point. Considering the influence of line impedance and load fluctuation factors, the dynamic weighted coefficient method is used to express the similarity or difference measurement of the operating data between the two measurement points. The formula is expressed as follows:

[0044]

[0045] Among them, S uv is the difference measure, μ u is the mean voltage of the power supply measuring point u, μ v is the mean voltage of the power supply measuring point v, σ t is the set voltage difference tolerance threshold, w1 and w2 are similarity measurement weights and w1>w2, e is the base of the natural logarithm, Pearson (U u , U v ) is the Pearson correlation coefficient between the power supply measurement point u and the power supply measurement point v. When the similarity measure is greater than 0.95, it is determined that the power supply measurement point u and the power supply measurement point v are in the same branch. The fundamental wave phase angles of u and v are selected. When the absolute value of the phase difference Δθ=|θ u -θ v |>5°, then a hierarchical connection relationship u→v is established, where θ u is the fundamental wave phase angle of the power supply measuring point u, θ v is the fundamental wave phase angle of the power supply measuring point v.

[0046] In an embodiment of the present application, the construction of a mapping relationship of power consumption information characteristics reflecting the power supply topology structure includes selecting a power supply measurement point as a root node, using a recursive algorithm to search for all adjacent nodes of the power supply node according to the connection relationship, and expanding to all measurement points layer by layer to construct a hierarchical structure and spatiotemporal distribution network model based on the power supply topology.

[0047] In an optional embodiment, the construction of a mapping relationship of power consumption information characteristics reflecting the power supply topology structure includes obtaining voltage, current, and power flow data of each measuring point in the power supply area, pre-setting fixed weights for the similarity measurement between the measuring points, calculating the cosine similarity between the voltage and current data vectors of the measuring points u and v, and if the cosine similarity is greater than a preset threshold of 0.9, determining that the measuring points u and v are in the same branch; directly establishing a hierarchical connection based on whether the absolute value of the phase difference exceeds 5° without recursive expansion, and using a breadth-first search (BFS) to traverse all measuring points to generate a linear topology structure.

[0048] In another optional embodiment, the construction of a mapping relationship of power consumption information characteristics reflecting the power supply topology structure includes obtaining voltage and current data of each measuring point, presetting fixed weights, ignoring the influence of line impedance and load fluctuations, and directly calculating the Pearson correlation coefficient of measuring points u and v. If the coefficient is greater than 0.9, it is determined to be the same branch. Starting from any measuring point, a one-way traversal is used to connect all measuring points with qualified correlation coefficients to form a single-layer topology structure.

[0049] S2: Determine the power supply scenario type based on the hierarchical architecture of the power supply topology and the temporal and spatial distribution characteristics of the load, and extract the power supply characteristics and power consumption patterns.

[0050] Furthermore, according to the corresponding hierarchical architecture and the distribution characteristics of different loads in time and space, a network model based on the power supply topology relationship is constructed and the depth of the network model, the branch node density and global density of the layer where each measuring point is located are determined. When the depth of the network model reaches a certain level and the branch node density is high, it can be determined as a centralized power supply or intensive load scenario; otherwise, it can be determined as a sparse power supply or dispersed load scenario; according to the power supply topology network model and its depth, the power supply characteristics and power consumption patterns corresponding to the power supply scenario are determined.

[0051] Furthermore, the depth L of the network model is determined, the top node is defined as layer 0, and the branch node density and global density of the power supply measurement point k in layer l are calculated. The formula is expressed as:

[0052]

[0053] Among them, D l(k) is the branch node density of the lth layer, D is the global density, B l(k) is the number of branch nodes in the lth layer where the power supply measurement point k is located in the network model, N l(k) is the total number of all nodes in the lth layer where the power supply measurement point k is located in the network model, B is the total number of all nodes with branches in the network model, N is the number of all nodes in the network model, and k is the power supply measurement point;

[0054] In the implementation mode of the present application, the scenario type is determined by combining the operating parameters of the power supply measurement point and the network model using the following rules: if L ≥ 4 and D > 0.7, it is determined to be a centralized power supply or intensive load scenario; otherwise, it can be determined to be a sparse power supply or dispersed load scenario. Based on the power supply topology network model and its depth, the power supply characteristics and power consumption patterns in the scenario are summarized.

[0055] In an optional embodiment, determining the scenario type includes constructing a tree topology model based on the hierarchical connection relationship between the measuring points, and only recording the network depth L. If the network depth L≥4, it is determined to be a load-intensive scenario; if L<4, it is determined to be a load-distributed scenario, and the preset power consumption pattern is allocated according to the scenario type.

[0056] In another optional implementation, determining the scenario type includes counting the total number N of nodes in the power supply topology network, and determining it as a load-intensive scenario if N>100, otherwise determining it as a distributed scenario.

[0057] S3: Establish a feature optimization model based on the power supply scenario, dynamically adjust the feature weights and combine physical constraints with weight threshold limits to implement a double screening mechanism and generate a cross-fusion feature set.

[0058] Furthermore, electricity usage features were classified into two categories, and appropriate methods were used to evaluate the importance of different features. A power usage feature optimization model was then established based on power usage scenarios. During the iterative optimization process, a maximum number of iterations was set, and feature weights were dynamically adjusted at each iteration until convergence conditions were met. Subsequently, a dual feature selection and optimization process was performed based on the physical constraints of electricity usage information and feature weight thresholds, eliminating marginal features and constructing a key feature space for electricity usage information.

[0059] Furthermore, for character features, the chi-square test and CatBoost method are used to evaluate the importance of each character feature, and the feature correlation degree FCD is the feature importance score;

[0060] For numerical features, the feature correlation degree FCD is expressed as:

[0061]

[0062] in, is the arithmetic mean of the mth eigenvalues of all power supply measurement points, is the arithmetic mean of the branch node density of all levels, Q m represents the feature importance score, k is the test point, D l(k) is the branch node density of the lth layer, m is the numerical feature, n is the total number of test points, FCD m is the feature correlation.

[0063] The weight of the mth feature is expressed as:

[0064] W m =α*FCD m +β*FCD m

[0065] Among them, W m is the weight of the mth feature, α and β are the scene type adjustment coefficients, m is a numerical feature, FCD m is the feature correlation.

[0066] Set the maximum number of iterations t and perform the following steps until the convergence condition is met: in the i-th iteration, calculate the feature weight set and the difference between adjacent iteration weights. When the difference between iteration weights is less than 0.1, terminate the iteration. Otherwise, dynamically update the complexity penalty coefficient of the feature splitting criterion and the scene type adjustment coefficient.

[0067] Furthermore, based on the physical constraints of electricity information and the feature weight threshold limit, a double screening optimization is performed to propose edge features. The first screening condition is: the feature weight value W m <0.05; Second screening condition: When the feature data violates the physical constraints of electricity consumption information and the error exceeds 5%, the features that meet the conditions at the same time will be eliminated to retain the cross-fusion features of power supply scenarios and topological relationships.

[0068] S4: Inject confusion variables into the fusion feature set to form an extended feature library, and evaluate the optimization effect based on the confusion resistance index. If the result is not up to standard, iterative optimization is triggered.

[0069] Acquire environmental variables related to power consumption behavior that are unrelated to the power supply topology, build a confusion variable library and mix them in proportion to form an extended feature set. Evaluate the dynamic weights of all features through the feature weight calculation method. According to the confusion resistance index (CRI) feature optimization effect, if the target is not achieved, trigger the feature optimization model, and continue to iterate, calculate, and screen to improve the features.

[0070] Furthermore, we obtain environmental variables related to power consumption behavior that are unrelated to the power supply topology to form confounding variables. We inject the confounding variables into the cross-fusion features at a ratio of 7:3 to generate an extended feature set. We evaluate the dynamic weights of all features through the feature optimization model, and record the number of retained confounding variables and other values. We use the Confusion Resistance Index (CRI) as an evaluation indicator. The Confusion Resistance Index is expressed as:

[0071]

[0072] Among them, CRI is the confusion resistance index, RSC is the number of retained cross-fusion features, OSC is the number of original cross-fusion features, CE is the number of correctly eliminated confounding variables, CS is the total number of injected confounding variables, and SF is the total number of features; when CRI>1.5, it indicates that the optimization effect is good. When CRI≤1.5, the feature optimization model is triggered, and iteration continues to readjust the model and improve the model parameters.

[0073] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

[0074] Example 2, reference Figure 2 , which is the second embodiment of the present invention, provides an electricity consumption information feature optimization system that takes into account power supply scenarios and topological relationships, including a data acquisition and preprocessing module, a power supply scenario identification and feature extraction module, a feature optimization model construction and screening module, and an extended feature library generation and optimization effect evaluation module.

[0075] The data acquisition and preprocessing module includes a measuring point electricity consumption information acquisition unit and an operation data correlation analysis unit, which are used to obtain electricity consumption information data of each measuring point in the power supply area, calculate the operation data correlation between the measuring points through a similarity measurement method, and construct a power consumption information characteristic mapping relationship reflecting the power supply topology structure.

[0076] The power supply scenario identification and feature extraction module includes a power supply topology analysis unit, a scenario type determination unit, and a power supply characteristic and power consumption pattern extraction unit. It is used to determine the power supply scenario type based on the hierarchical architecture of the power supply topology and the temporal and spatial distribution characteristics of the load, extract the power supply characteristics and power consumption patterns related to the power supply scenario, and assist in scenario type determination through network depth, branch node density and global density indicators.

[0077] The feature optimization model construction and screening module includes a feature classification unit, a feature importance evaluation unit, a feature weight dynamic adjustment unit, and a dual screening mechanism execution unit, which is used to divide the electricity consumption information features into two categories: character type and numerical type, and evaluate the feature importance through chi-square test and normalized covariance correlation respectively, dynamically update the feature weight according to the scenario type adjustment coefficient, set the convergence condition to control the number of iterations, and execute the dual screening mechanism.

[0078] The extended feature library generation and optimization effect evaluation module includes a confusion variable injection unit, an extended feature library generation unit, a confusion resistance index calculation unit, and an iterative optimization triggering unit. It is used to inject irrelevant environmental variables as confusion variables into the cross-fusion feature set to generate an extended feature library, evaluate the retention of confounding variables through feature weight calculation, and quantify the optimization effect based on the confusion resistance index formula.

[0079] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

[0080] Embodiment 3, the third embodiment of the present invention, is different from the first two embodiments in that:

[0081] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0082] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0083] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering, or processing in another suitable manner as necessary, and then stored in a computer memory.

[0084] It should be understood that various parts of the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application-specific integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0085] Example 4 is the fourth embodiment of the present invention, which provides a method for optimizing power consumption information characteristics considering power supply scenarios and topological relationships. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through experiments.

[0086] 50,000 pieces of electricity usage information from required measurement points within the power supply area were obtained. Existing technologies and the method of the present invention were applied to ensure data relevance and accuracy. The power supply measurement point data were input into the fault location model and the electricity theft detection model using the respective technologies. The existing technology did not consider scenario adaptation and performed feature dimensionality reduction. The data processing speed and dimensionality reduction of the two technologies were recorded, as well as the results of their application in terms of fault location and electricity theft risk detection accuracy.

[0087]

[0088] Correspondingly, an electricity consumption information database containing 15 cross-fusion features was used, and 4 confounding variables were injected in proportion to form an extended feature library. After double feature screening, 14 cross-fusion features were retained, and the confounding variable 4 was correctly eliminated. The calculated confusion resistance index CRI was close to the integer 2, which further demonstrated the effectiveness of this optimization method.

[0089] Clearly, the electricity theft detection algorithm of the present invention outperforms existing technologies in terms of data processing speed, feature dimensionality, and application effectiveness. The present invention fully considers power supply operation patterns, multi-scenario adaptation, and dynamic adaptability, achieving cross-fusion feature selection of electricity usage information. It fully utilizes the underlying rules and constraints of data information scenarios to improve the accuracy of data information analysis and processing applications, forming optimized data information features that reflect physical properties. This helps to enhance feature value and model performance through elimination, screening, and interactive fusion, providing better data support for power supply operation management services.

Claims

1. A method for optimizing power consumption information characteristics considering power supply scenarios and topological relationships, characterized by: include, Obtain power consumption data from measurement points within the power supply area, and build a mapping relationship of power consumption characteristics that reflects the power supply topology based on the similarities and differences between the operating data and configuration information between the measurement points; Determine the power supply scenario type based on the hierarchical architecture of the power supply topology and the temporal and spatial distribution characteristics of the load, and extract the power supply characteristics and power consumption patterns; Establish a feature optimization model based on power supply scenarios. By dynamically adjusting feature weights and combining physical constraints with weight threshold limits, a double screening mechanism is implemented to generate a cross-fusion feature set. Confusion variables are injected into the fusion feature set to form an extended feature library, and the optimization effect is evaluated based on the confusion resistance index. If the index is not met, iterative optimization is triggered.

2. The method for optimizing power consumption information characteristics considering power supply scenarios and topological relationships according to claim 1, characterized in that: The electricity consumption information data of the measuring points includes calculating the voltage, current and power flow between the measuring points; The operation data correlation includes calculating the operation data correlation by a similarity measurement method between measurement points; The method of constructing a mapping relationship of power consumption information characteristics reflecting the power supply topology structure includes: a similarity measurement method between measurement points including a dynamic weighted analysis combining line impedance and load fluctuation factors, and expanding the connection relationship of measurement points layer by layer based on a recursive algorithm to construct a power supply topology network model reflecting the hierarchical structure and spatiotemporal distribution.

3. The method for optimizing power consumption information characteristics considering power supply scenarios and topological relationships according to claim 2, characterized in that: The dynamic weighted analysis includes linearly weighting the exponential decay term of the voltage mean difference between the measuring points with the Pearson correlation coefficient. When the weighted similarity metric exceeds a preset threshold, the measuring points are determined to be located in the same branch, and a hierarchical connection relationship is established based on the fundamental wave phase difference. Among them, the weight of the voltage mean difference term is smaller than the weight of the correlation coefficient term.

4. The method for optimizing power consumption information characteristics considering power supply scenarios and topological relationships according to claim 3, characterized in that: The extraction of power supply characteristics and power consumption patterns includes determining the scenario type based on the depth, branch node density, and global density of the power supply topology network model. When the network depth reaches a preset level and the branch node density exceeds a threshold, it is determined to be a load-intensive scenario; otherwise, it is determined to be a load-distributed scenario. Among them, the branch node density is calculated by the ratio of the number of branch nodes in each layer to the total number of nodes, and the global density is calculated by the ratio of the number of branch nodes in the entire network to the total number of nodes.

5. The method for optimizing power consumption information characteristics considering power supply scenarios and topological relationships according to claim 4, characterized in that: The establishment of a feature optimization model based on power supply scenarios includes classifying power consumption information features into character and numeric types, evaluating feature importance through chi-square tests and normalized covariance correlation, dynamically updating feature weights based on scenario type adjustment coefficients, and setting convergence conditions to control the number of iterations until the difference in weights between adjacent iterations is lower than a preset threshold. Among them, the normalized covariance correlation includes calculating the covariance of the numerical features of the measurement points and the branch node density of the hierarchy, and generating the final weight in combination with the feature importance score.

6. The method for optimizing power consumption information characteristics considering power supply scenarios and topological relationships according to claim 5, characterized in that: The implementation of the double screening mechanism includes setting the first screening condition as the feature weight being lower than a preset threshold, and the second screening condition as the feature data violating physical constraints and the error exceeding the tolerance range, retaining features that meet both the first and second screening conditions, and eliminating redundant and marginal features.

7. The method for optimizing power consumption information characteristics considering power supply scenarios and topological relationships according to claim 6, characterized in that: The forming of the extended feature library includes injecting irrelevant environmental variables as confounding variables into the cross-fusion feature set in proportion to generate the extended feature library, evaluating the retention of the confounding variables through feature weight calculation, and quantifying the optimization effect based on the confounding resistance index formula; The formula for the confusion resistance index is: Among them, CRI is the confounding resistance index, RSC is the number of retained cross-fusion features, OSC is the number of original cross-fusion features, CE is the number of confounding variables correctly eliminated, CS is the total number of injected confounding variables, and SF is the total number of features. When CRI>1.5, it indicates that the optimization effect is good. When CRI≤1.5, the feature optimization model is triggered and it continues to iterate, readjust the model and improve the model parameters. The iterative optimization includes adjusting the complexity penalty coefficient and the scene type adjustment coefficient of the feature weight model according to the confusion resistance index result, and re-executing the feature weight calculation and double screening until the confusion resistance index reaches a preset target value.

8. A system using the method for optimizing power consumption information characteristics considering power supply scenarios and topological relationships according to any one of claims 1 to 7, characterized in that: It includes data acquisition and preprocessing module, power supply scenario recognition and feature extraction module, feature optimization model construction and screening module, and extended feature library generation and optimization effect evaluation module; The data acquisition and preprocessing module includes a measurement point electricity consumption information acquisition unit and an operation data correlation analysis unit, which are used to obtain electricity consumption information data of each measurement point in the power supply area, calculate the correlation of operation data between measurement points through a similarity measurement method, and construct a mapping relationship of electricity consumption characteristics reflecting the power supply topology structure; The power supply scenario identification and feature extraction module includes a power supply topology analysis unit, a scenario type determination unit, and a power supply characteristic and power consumption pattern extraction unit. It is used to determine the power supply scenario type based on the hierarchical structure of the power supply topology and the spatiotemporal distribution characteristics of the load, extract the power supply characteristics and power consumption patterns related to the power supply scenario, and assist in scenario type determination through network depth, branch node density, and global density indicators; The feature optimization model construction and screening module includes a feature classification unit, a feature importance evaluation unit, a feature weight dynamic adjustment unit, and a dual screening mechanism execution unit, which is used to classify electricity consumption information features into two categories: character type and numerical type, and evaluate feature importance through chi-square test and normalized covariance correlation respectively, dynamically update feature weights according to scenario type adjustment coefficients, set convergence conditions to control the number of iterations, and execute the dual screening mechanism; The extended feature library generation and optimization effect evaluation module includes a confusion variable injection unit, an extended feature library generation unit, a confusion resistance index calculation unit, and an iterative optimization triggering unit. It is used to inject irrelevant environmental variables as confusion variables into the cross-fusion feature set to generate an extended feature library, evaluate the retention of confounding variables through feature weight calculation, and quantify the optimization effect based on the confusion resistance index formula.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method for optimizing power consumption information characteristics considering power supply scenarios and topological relationships described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for optimizing power consumption information characteristics considering power supply scenarios and topological relationships described in any one of claims 1 to 7 are implemented.

Images