Partition optimization method and system for power distribution network

Abstract

The invention provides a power distribution network partition optimization method and system. The method comprises the steps that a non-inductive load characteristic index set is constructed, non-inductive load characteristic indexes are constructed into characteristic vectors, electrical attributes of a power supply unit are constructed into attribute vectors, and then a two-dimensional evaluation model is constructed; obtaining a plurality of candidate partitions based on electrical attributes of the power supply units, calculating a comprehensive evaluation value of each load point and each power supply unit through a two-dimensional evaluation model, distributing each load point to the candidate partitions meeting a preset partition requirement through the comprehensive evaluation value, obtaining a preliminary partition scheme, constructing a global optimization model, and obtaining a final partition scheme; and optimizing each candidate partition to obtain a preliminary partition optimization scheme, constructing a corresponding power flow balance equation, and locally adjusting the preliminary partition optimization scheme to obtain a power distribution network partition optimization scheme so as to optimize the power distribution network partition. According to the invention, multi-dimensional evaluation is carried out on the power distribution network to optimize the reasonability of load division of the power distribution network, so that the operation reliability of the power grid is improved.

Description

A method and system for optimizing power distribution network zoning Technical Field

[0001] This invention relates to the field of intelligent planning and operation optimization technology for power systems, and in particular to a method and system for zoning optimization of distribution networks. Background Technology

[0002] With the acceleration of urbanization and the continuous development of new power systems, the load scale of distribution networks continues to climb, and the load structure is becoming increasingly complex. After diverse users such as residents, businesses, industries, and new infrastructure are connected to the power grid, they will exhibit unbalanced, multi-scale, highly time-varying, and highly uncertain operating characteristics. Therefore, a large number of non-impact load resources with adjustable capabilities but without affecting user perception are constantly being connected, posing a new challenge to traditional distribution network planning and zoning methods.

[0003] Currently, the existing distribution network zoning methods mainly rely on topology and electrical parameters. Although some studies have introduced load factors, they often only divide the network based on load size or typical curves. In other words, the existing technical means lack a systematic characterization of the multi-dimensional characteristics of non-inductive loads, such as peak-valley characteristics, fluctuation degree, regulation potential, and response speed. It is difficult to accurately capture the differences and complementary relationships between loads in different areas, resulting in significant deficiencies in load balance and capacity utilization in the zoning results. Furthermore, it is prone to load concentration and capacity overruns, leading to unreliable power grid operation. Summary of the Invention

[0004] To address the aforementioned issues, this invention proposes a distribution network zoning optimization method and system, which enables multi-dimensional evaluation of the distribution network to optimize the rationality of distribution network zoning load allocation, thereby improving the reliability of power grid operation.

[0005] To achieve the above objectives, embodiments of the present invention provide a distribution network zoning optimization method, comprising: constructing a set of non-inductive load characteristic indicators based on pre-acquired original load point data; constructing the non-inductive load characteristic indicators as feature vectors and constructing the pre-acquired electrical attributes of power supply units as attribute vectors; constructing a two-dimensional evaluation model based on the feature vectors and attribute vectors; obtaining several candidate zones based on the pre-acquired electrical attributes of power supply units, calculating the comprehensive evaluation value of each load point and each power supply unit through the two-dimensional evaluation model, and allocating each load point to a candidate zone that meets the preset zoning requirements based on the comprehensive evaluation value, thereby obtaining a preliminary zoning scheme; constructing a global optimization model based on the preliminary zoning scheme, optimizing each candidate zone, thereby obtaining a preliminary zoning optimization scheme; constructing a corresponding power flow balance equation based on the preliminary zoning optimization scheme, and locally adjusting the preliminary zoning optimization scheme to obtain a distribution network zoning optimization scheme, thereby optimizing the distribution network zoning.

[0006] This invention proposes a distribution network zoning optimization method. It extracts non-inductive load characteristic indicators from raw load point data, constructs these indicators as feature vectors, and then constructs attribute vectors from pre-acquired electrical attributes of power supply units. This fully considers the flexibility and adjustability of the loads, as well as the rigidity and structural parameters of the power grid, providing a reliable data foundation for subsequent optimization analysis. Then, a two-dimensional evaluation model is constructed based on the feature vectors and attribute vectors to comprehensively evaluate each load point and each power supply unit, generating a preliminary zoning scheme. The preliminary zoning scheme is then globally optimized to obtain a preliminary optimized zoning scheme, effectively preventing load concentration and capacity overruns. Finally, a power flow balance equation is constructed to locally adjust the preliminary optimized zoning scheme to generate a distribution network zoning optimization scheme for further optimization. Therefore, by introducing a two-dimensional evaluation model and combining non-inductive load characteristics with the electrical attributes of power supply units, the load balance of the distribution network and the overall engineering feasibility of the scheme can be effectively improved. This ensures that the zoning results are not only theoretically optimal but also meet the safety and stability requirements of the power grid in practical applications, thereby improving the reliability of power grid operation.

[0007] Furthermore, based on the pre-acquired raw load point data, a set of non-intrusive load characteristic indicators is constructed, including: collecting the spatial coordinates of load points, predicting annual load data and typical daily load curves to obtain raw load point data; constructing a raw daily load curve matrix based on the raw load point data; preprocessing the raw daily load curve matrix and extracting several characteristic indicators from the raw daily load curve matrix to obtain a set of non-intrusive load characteristic indicators.

[0008] In the above scheme, the spatial coordinates of load points, predicted annual load data and typical daily load curves are collected and an original daily load curve matrix is ​​constructed. Then, several characteristic indicators are extracted from the original daily load curve matrix, which can comprehensively quantify load characteristics and provide a reliable data foundation for the construction of subsequent evaluation models. This ensures that the zoning results are not only theoretically optimal, but also meet the safety and stability requirements of the power grid in practical applications, thereby improving the reliability of power grid operation.

[0009] Furthermore, the non-inductive load characteristic index is constructed as a feature vector, and the pre-acquired electrical attributes of the power supply unit are constructed as an attribute vector, including: calculating the normalized value of the non-inductive load characteristic index based on the original value of the non-inductive load characteristic index, constructing the feature vector based on the normalized value of the non-inductive load characteristic index; acquiring the topology, electrical parameters, and power supply boundary of the power supply unit; and constructing the attribute vector based on the topology, electrical parameters, power supply boundary, and attribute dimension of the power supply unit.

[0010] In the above scheme, the characteristics of non-inductive loads and the electrical attributes of power supply units are constructed as feature vectors and attribute vectors, respectively. The complex load and power grid entities are abstracted into vectors in mathematical space, which effectively eliminates the dominant influence of different dimensions on the calculation results, provides a reliable data foundation for the construction of subsequent evaluation models, and ensures that the zoning results are not only theoretically optimal, but also meet the safety and stability requirements of the power grid in practical applications, thereby improving the reliability of power grid operation.

[0011] Furthermore, a two-dimensional evaluation model is constructed based on feature vectors and attribute vectors, including: calculating the difference measure of non-inductive load characteristics based on feature vectors, the first weight of non-inductive load characteristic indicators, and preset cluster centers; calculating the difference measure of power supply unit electrical attributes based on attribute vectors, the second weight of power supply unit electrical attributes, and preset cluster centers; summing the difference measure of non-inductive load characteristics and the difference measure of power supply unit electrical attributes, and setting weight coefficients in the corresponding dimensions of the difference measure of non-inductive load characteristics and the difference measure of power supply unit electrical attributes to obtain the two-dimensional evaluation model.

[0012] In the above scheme, the difference in characteristics of inductive loads and the difference in electrical attributes of power supply units are calculated using feature vectors and attribute vectors, respectively. Adjustable weight coefficients are introduced in the calculation process to establish a two-dimensional evaluation model that considers the similarity and complementarity of load characteristics within the region and the electrical carrying capacity and boundaries of the power supply unit itself. The two-dimensional evaluation model outputs a zoning scheme that meets the needs of specific scenarios, effectively preventing the problems of load concentration and capacity overrun. It can ensure that the zoning results are not only theoretically optimal, but also meet the safety and stability requirements of the power grid in practical applications, thereby improving the reliability of power grid operation.

[0013] Furthermore, based on the pre-acquired electrical attributes of the power supply units, several candidate zones are obtained. A two-dimensional evaluation model is used to calculate the comprehensive evaluation value of each load point and each power supply unit. Based on the comprehensive evaluation value, each load point is assigned to a candidate zone that meets the preset zone requirements, resulting in a preliminary zone allocation scheme. This includes: based on the pre-acquired electrical attributes of the power supply units, a preliminary combination of each load point and its corresponding power supply unit is performed to obtain several candidate zones; based on the two-dimensional evaluation model, the difference in inductive load characteristics and the difference in electrical attributes of each power supply unit between each load point and each power supply unit are calculated to obtain the comprehensive evaluation value of each load point and each power supply unit; a comprehensive evaluation matrix is ​​constructed based on the comprehensive evaluation value of each load point and each power supply unit; and based on the comprehensive evaluation matrix, each load point is assigned to a candidate zone that meets the preset zone requirements, resulting in a preliminary zone allocation scheme.

[0014] In the above scheme, each load point and its corresponding power supply unit are initially combined based on the electrical attributes of the power supply unit to generate several candidate partitions. Then, a two-dimensional evaluation model is used to calculate the difference in inductive load characteristics and the difference in electrical attributes of each power supply unit between each load point and each power supply unit, thereby obtaining a comprehensive evaluation value for each load point and each power supply unit. A comprehensive evaluation matrix is ​​then constructed to allocate each load point to a candidate partition that meets the preset partitioning requirements, resulting in a preliminary partitioning scheme. Thus, each load point is assigned to the region that best matches its characteristics and is most compatible with the attributes of the power supply unit under the given conditions. This ensures that the generated preliminary partitioning scheme achieves a relatively optimal state in a localized area, effectively preventing load concentration and capacity overruns. It ensures that the partitioning result is not only theoretically optimal but also meets the safety and stability requirements of the power grid in practical applications, improving the reliability of power grid operation.

[0015] Furthermore, a global optimization model is constructed based on the preliminary partitioning scheme to optimize each candidate partition and obtain a preliminary partitioning optimization scheme. This includes: constructing a global optimization model based on the preliminary partitioning scheme and a preset objective function; optimizing the attribution relationship between each load point and each candidate partition until the preset global partitioning optimization requirements are met, thus obtaining a preliminary partitioning optimization scheme.

[0016] In the above scheme, a global optimization model is constructed from the preliminary zoning scheme and a preset objective function. The relationship between each load point and each candidate zone is optimized, and the load balance between zones and the deep complementarity of loads within the region, which were ignored in the preliminary zoning scheme, are optimized. This effectively prevents the problems of load concentration and capacity overrun. It ensures that the zoning results are not only theoretically optimal, but also meet the safety and stability requirements of the power grid in practical applications, thereby improving the reliability of power grid operation.

[0017] Furthermore, based on the preliminary zoning optimization scheme, corresponding power flow balance equations are constructed, and the preliminary zoning optimization scheme is locally adjusted to obtain a distribution network zoning optimization scheme for optimizing the distribution network zoning. This includes: performing preliminary zoning of the distribution network based on the preliminary zoning optimization scheme to obtain zoning results; constructing power flow balance equations corresponding to each zoning result, and mapping load points to corresponding power flow nodes according to the power flow balance equations; performing power flow verification on each power flow node; if the power flow verification result does not meet the preset operating requirements, then locally adjusting the preliminary zoning optimization scheme until the power flow verification result meets the preset operating requirements, and constructing an operating optimization model to perform operating optimization on the preliminary zoning optimization scheme that meets the preset operating requirements to obtain a distribution network zoning optimization scheme for optimizing the distribution network zoning.

[0018] In the above scheme, the partitioning results are obtained from the preliminary partitioning optimization scheme, the power flow balance equations corresponding to each partitioning result are constructed, and the load points are mapped to the corresponding power flow nodes. Finally, the preliminary partitioning optimization scheme is optimized in operation, and then the preliminary partitioning optimization scheme is further adjusted locally. The introduction of a power flow verification mechanism effectively prevents the problems of load concentration and capacity over-limit. It can ensure that the partitioning results are not only theoretically optimal, but also meet the safety and stability requirements of the power grid in practical applications, thereby improving the reliability of power grid operation.

[0019] This invention also provides a distribution network zoning optimization system, including: a characteristic index set construction module, a vector construction module, a two-dimensional evaluation model construction module, a preliminary zoning scheme acquisition module, a preliminary zoning optimization scheme acquisition module, and a distribution network zoning optimization module; the characteristic index set construction module is used to construct a set of non-inductive load characteristic indices based on pre-acquired raw load point data; the vector construction module is used to construct feature vectors from the non-inductive load characteristic indices and attribute vectors from the pre-acquired electrical attributes of power supply units; the two-dimensional evaluation model construction module is used to construct a two-dimensional evaluation model based on the feature vectors and attribute vectors; the preliminary zoning scheme acquisition module... The system is used to obtain several candidate partitions based on the pre-acquired electrical attributes of power supply units, calculate the comprehensive evaluation value of each load point and each power supply unit through a two-dimensional evaluation model, and allocate each load point to a candidate partition that meets the preset partitioning requirements based on the comprehensive evaluation value, thus obtaining a preliminary partitioning scheme; the preliminary partitioning optimization scheme acquisition module is used to build a global optimization model based on the preliminary partitioning scheme, optimize each candidate partition, and obtain a preliminary partitioning optimization scheme; the distribution network partitioning optimization module is used to build the corresponding power flow balance equation based on the preliminary partitioning optimization scheme, and locally adjust the preliminary partitioning optimization scheme to obtain a distribution network partitioning optimization scheme, so as to optimize the distribution network partitioning.

[0020] This invention proposes a distribution network zoning optimization system. It extracts non-inductive load characteristic indicators from raw load point data, constructs these indicators as feature vectors, and then constructs attribute vectors from pre-acquired electrical attributes of power supply units. This fully considers the flexibility and adjustability of the loads, as well as the rigidity and structural parameters of the power grid, providing a reliable data foundation for subsequent optimization analysis. A two-dimensional evaluation model is then constructed based on the feature vectors and attribute vectors to comprehensively evaluate each load point and each power supply unit, generating a preliminary zoning scheme. This preliminary zoning scheme is then globally optimized to obtain a preliminary optimized zoning scheme, effectively preventing load concentration and capacity overruns. Finally, a power flow balance equation is constructed to locally adjust the preliminary optimized zoning scheme to generate a final optimized distribution network zoning scheme. Therefore, by introducing a two-dimensional evaluation model and combining non-inductive load characteristics with the electrical attributes of power supply units, the system effectively improves the load balance of the distribution network and the overall engineering feasibility of the scheme. This ensures that the zoning results are not only theoretically optimal but also meet the safety and stability requirements of the power grid in practical applications, thereby improving the reliability of power grid operation.

[0021] Furthermore, the characteristic index set construction module is used to construct a set of non-intrusive load characteristic indicators based on the pre-acquired raw load point data. It includes: a first data acquisition unit, a matrix construction unit, and a characteristic index extraction unit. The data acquisition unit is used to collect the spatial coordinates of the load points, the predicted annual load data, and the typical daily load curves to obtain the raw load point data. The matrix construction unit is used to construct the raw daily load curve matrix based on the raw load point data. The characteristic index extraction unit is used to preprocess the raw daily load curve matrix and extract several characteristic indicators from the raw daily load curve matrix to obtain the set of non-intrusive load characteristic indicators.

[0022] Furthermore, the vector construction module is used to construct feature vectors from the non-inductive load characteristic indicators and to construct attribute vectors from the pre-acquired electrical attributes of the power supply unit. This includes a feature vector construction unit, a second data acquisition unit, and an attribute vector construction unit. The feature vector construction unit is used to calculate the normalized values ​​of the non-inductive load characteristic indicators based on their original values, and to construct feature vectors based on these normalized values. The second data acquisition unit is used to acquire the topology, electrical parameters, and power supply boundaries of the power supply unit. The attribute vector construction unit is used to construct attribute vectors based on the topology, electrical parameters, power supply boundaries, and attribute dimension of the power supply unit. Attached Figure Description

[0023] Figure 1 is a flowchart illustrating the steps of a distribution network zoning optimization method according to a certain embodiment of the present invention; Figure 2 is an example of the execution flow of a distribution network zoning optimization method according to a certain embodiment of the present invention; Figure 3 is a schematic diagram of the module structure of a distribution network zoning optimization system according to a certain embodiment of the present invention. Detailed Implementation

[0024] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] Example 1: Refer to Figure 1. Figure 1 is a flowchart illustrating the steps of a power distribution network zoning optimization method provided in a certain embodiment of the present invention. As shown in Figure 1, this embodiment of the invention proposes a distribution network zoning optimization method, including steps 101 to 106, each step being as follows: Step 101, constructing a set of non-inductive load characteristic indicators based on pre-acquired original load point data; Step 102, constructing feature vectors from the non-inductive load characteristic indicators and attribute vectors from the pre-acquired power supply unit electrical attributes; Step 103, constructing a two-dimensional evaluation model based on the feature vectors and attribute vectors; Step 104, obtaining several candidate zones based on the pre-acquired power supply unit electrical attributes, calculating the comprehensive evaluation value of each load point and each power supply unit through the two-dimensional evaluation model, and allocating each load point to a candidate zone that meets the preset zoning requirements based on the comprehensive evaluation value, thus obtaining a preliminary zoning scheme; Step 105, constructing a global optimization model based on the preliminary zoning scheme, optimizing each candidate zone, thus obtaining a preliminary zoning optimization scheme; Step 106, constructing the corresponding power flow balance equation based on the preliminary zoning optimization scheme, and locally adjusting the preliminary zoning optimization scheme to obtain a distribution network zoning optimization scheme, thereby optimizing the distribution network zoning.

[0026] One possible implementation method is illustrated in Figure 2, which is a schematic diagram of an example execution flow of a distribution network zoning optimization method provided by a certain embodiment of the present invention. As shown in Figure 2, taking distribution network zoning optimization considering non-inductive loads as an example, before distribution network zoning optimization, it is necessary to collect data and extract characteristics of various typical non-inductive loads, that is, to collect original load point data and extract non-inductive load characteristic indicators to construct a set of non-inductive load characteristic indicators. After completing the extraction of non-inductive load characteristic indicators, the electrical attributes of the distribution network power supply unit are further combined to construct feature vectors of non-inductive load characteristic indicators and attribute vectors of power supply unit electrical attributes. Then, a two-dimensional evaluation model is constructed using feature vectors and attribute vectors. In this embodiment, the two-dimensional evaluation model is explained as a non-inductive load characteristic-power supply unit two-dimensional evaluation model, which will not be elaborated further below. Based on the construction of a dual-dimensional evaluation model of inductive load characteristics and power supply units, each load point is assigned to a candidate partition that meets the preset partitioning requirements through comprehensive evaluation values. In this embodiment, the preset partitioning requirement is the highest matching degree with the load point. That is, each load point is assigned to the candidate partition with the highest matching degree with the load point through comprehensive evaluation values, thereby obtaining a preliminary partitioning scheme. Then, a global optimization model is constructed. Taking the preliminary partitioning as input, an objective function is established to optimize each candidate partition, resulting in a preliminary partitioning optimization scheme. Finally, after completing the partitioning optimization, the partitioning results are obtained according to the preliminary partitioning optimization scheme, and the corresponding power flow balance equations are established. The operating status of each node is checked, and the preliminary partitioning optimization scheme is locally adjusted according to the check results, ultimately generating a distribution network partitioning optimization scheme, thereby achieving the optimization of the distribution network partitioning.

[0027] This invention proposes a distribution network zoning optimization method. It extracts non-inductive load characteristic indicators from raw load point data, constructs these indicators as feature vectors, and then constructs attribute vectors from pre-acquired electrical attributes of power supply units. This fully considers the flexibility and adjustability of the loads, as well as the rigidity and structural parameters of the power grid, providing a reliable data foundation for subsequent optimization analysis. Then, a two-dimensional evaluation model is constructed based on the feature vectors and attribute vectors to comprehensively evaluate each load point and each power supply unit, generating a preliminary zoning scheme. The preliminary zoning scheme is then globally optimized to obtain a preliminary optimized zoning scheme, effectively preventing load concentration and capacity overruns. Finally, a power flow balance equation is constructed to locally adjust the preliminary optimized zoning scheme to generate a distribution network zoning optimization scheme for further optimization. Therefore, by introducing a two-dimensional evaluation model and combining non-inductive load characteristics with the electrical attributes of power supply units, the load balance of the distribution network and the overall engineering feasibility of the scheme can be effectively improved. This ensures that the zoning results are not only theoretically optimal but also meet the safety and stability requirements of the power grid in practical applications, thereby improving the reliability of power grid operation.

[0028] A preferred approach involves constructing a set of non-intrusive load characteristic indicators based on pre-acquired raw load point data. This includes: collecting the spatial coordinates of load points, predicted annual load data, and typical daily load curves to obtain raw load point data; constructing a raw daily load curve matrix based on the raw load point data; preprocessing the raw daily load curve matrix and extracting several characteristic indicators from the raw daily load curve matrix to obtain the set of non-intrusive load characteristic indicators.

[0029] One preferred implementation scheme is shown in Figure 2. During the data acquisition phase, the spatial coordinates, predicted annual load data, and typical daily load curves of all load points in the distribution network are collected as raw load point data. This raw load point data is required to cover various types of non-invisible load resources, including residential, commercial, industrial, and new infrastructure loads. The spatial coordinates of the load points can be used to determine their location within the power grid topology; the predicted annual load data can provide a scale benchmark for the load points; and the typical daily load curves can reflect the relative fluctuation characteristics of the load within a day. Then, using the predicted annual load data as a total constraint, the typical daily load curves are normalized and scaled to be combined with each load point corresponding to its spatial coordinates, resulting in a raw daily load curve matrix for each load point and each time point. The specific calculation method is as follows: ; ; In the formula, Raw daily load point data; For load point Spatial coordinates; For load point Forecast annual load data; The typical number of days in a year (usually 365); Category of load point i Typical daily load curves for (residential, commercial, industrial, and new infrastructure) at time The value of , Category of load point i Typical daily load curves for (residential, commercial, industrial, and new infrastructure) at time The value of ; The sampling time interval; This represents the total number of load points. This represents the total number of time periods within a day, such as 96 points / day; This is the original daily load curve matrix.

[0030] Then, the original curve data is subjected to denoising, normalization, and time-series alignment, which is represented as preprocessing in this embodiment to ensure the comparability and integrity of data at different load points. The specific calculation method is as follows: (1) Denoising preprocessing, in this embodiment, the moving average method is used, and the specific expression is: ;in, Let i be the denoised load value at load point i at time t; To smooth the window width, the value is determined based on the sampling interval of the load curve and the load type, and is generally between 2 and 6 sampling points; It is used to simplify formula expressions and has no special meaning.

[0031] (2) Normalization preprocessing: ;in, The normalized load; and For all time periods The maximum and minimum values ​​on.

[0032] (3) Timing alignment preprocessing: In this embodiment, linear interpolation is used, and the specific expression is as follows: ; ;in, For load point At any moment Alignment load value; , For the original sampling time, satisfying ; These are the interpolation coefficients.

[0033] Finally, let's look at the aligned daily load curve. Multi-dimensional indicators are extracted from them. In this embodiment, the multi-dimensional indicators include seven indicators: peak-to-valley ratio, load factor, volatility, ramp rate, correlation, complementarity and imperceptible load adjustment potential. Thus, an imperceptible load characteristic indicator set is constructed. The expressions of each indicator are as follows: (i) peak-to-valley ratio; In the formula, The peak-to-valley ratio of load i; and These represent the maximum and minimum values ​​for all time periods within a typical day.

[0034] (ii) Load factor, which is the ratio of average load to maximum load, reflects utilization rate and stability; ; ;in, For load point The average load; For load point Loading factor; This represents the total number of time periods within a day, such as 96 points / day; For load point At any moment Alignment load value.

[0035] (iii) Volatility, i.e., the amplitude of load fluctuation over time; In the formula, For load point Volatility indicators This represents the total number of time periods within a day, such as 96 points / day; For load point At any moment Alignment load value; For load point The average load.

[0036] (iv) Maximum gradeability; In the formula, The maximum ramp rate at load point i; For load point At any moment Alignment load value; For load point At any moment -1 alignment load value; This represents the sampling time interval.

[0037] (v) Load correlation and complementarity; ; In the formula, The correlation coefficient between load points i and j; It is a complementarity index; For load point At any moment Alignment load value; For load point At any moment Alignment load value; For load point The average load.

[0038] (vi) Imperceptible load regulation potential, i.e., the capacity that an imperceptible load can release without affecting the user experience; since different types of imperceptible loads can be uniformly modeled as "adjustable amplitude × sensitivity × capacity", the specific definition is as follows: In the formula, The total adjustable potential of load point i; Let i be the set of inductive load cells contained in load point i; The sensorless adjustable amplitude of unit u; The sensitivity coefficient of unit u; This refers to the rated capacity or operating capacity of unit u.

[0039] Among them, the seamless adjustable amplitude This parameter characterizes the maximum adjustable percentage under physical conditions or operational constraints. Its value is determined based on the operational characteristics of various load types and the user's acceptable range. For example, residential loads are determined by comfort zone constraints, with an adjustable range typically between 10% and 25%, more specifically, this translates to raising air conditioning temperatures by 1-2°C and delaying water heater heating by 0.5-1 hour. Commercial loads are determined by operational standards and energy-saving management, with an adjustable range typically between 10% and 25%, more specifically, this translates to reducing lighting by 10%-20% and raising central air conditioning temperatures by 1-2°C. Industrial loads are determined by combining process continuity and safety regulations; continuous processes have a lower adjustable range, typically 5%-10%, while intermittent equipment can reach 15%-30%. New infrastructure loads are determined by operational simulation and equipment control strategies, with an adjustable range of approximately 5%-40%, more specifically, this translates to 15%-25% for data center cooling systems, 5%-10% for IT loads, 10%-20% for 5G base stations, and 20%-40% for charging piles. Furthermore, the sensitivity coefficient... The sensitivity coefficient is used to correct the degree to which the adjustable range is acceptable and responsive to users in actual operation. Its value can be determined in the following ways: First, experimental measurement: Parameters of typical equipment are adjusted in a laboratory or field environment. More specifically, this involves increasing the set temperature of an air conditioner and delaying the heating time of a water heater, and measuring the ratio of the actual power reduction to the theoretical adjustable range to obtain the sensitivity. Second, on-site monitoring: Large amounts of user operation data are collected using smart meters and energy consumption monitoring systems. The proportion of actual load reduction to the theoretical range is statistically analyzed in demand response events to fit the sensitivity coefficient. Third, statistical modeling: Cluster analysis, large-sample regression, or probabilistic models are used to quantify the average response degree of different types of loads under different operating conditions. Fourth, pilot operation and standard reference: In power grid demand response or load management pilot projects, the actual response records of the loads are combined with industry standards to correct and determine the typical sensitivity range for different types of loads. With the above sensitivity coefficients... The method for determining the values ​​is as follows: the sensitivity of residential loads is usually 0.4-0.8, i.e., air conditioners 0.4-0.6 and water heaters 0.6-0.8; the sensitivity of commercial loads is generally 0.3-0.7, i.e., lighting 0.3-0.5 and central air conditioning 0.5-0.7; the sensitivity of industrial loads is affected by the interruptibility of the process, with continuous processes taking 0.1-0.3 and intermittent equipment taking 0.3-0.5; the sensitivity range of new infrastructure loads is about 0.1-0.7, i.e., data center IT load 0.1-0.3, cooling systems 0.5-0.7 and 5G base stations 0.3-0.5.

[0040] In the above scheme, the spatial coordinates of load points, predicted annual load data and typical daily load curves are collected and an original daily load curve matrix is ​​constructed. Then, several characteristic indicators are extracted from the original daily load curve matrix, which can comprehensively quantify load characteristics and provide a reliable data foundation for the construction of subsequent evaluation models. This ensures that the zoning results are not only theoretically optimal, but also meet the safety and stability requirements of the power grid in practical applications, thereby improving the reliability of power grid operation.

[0041] A preferred embodiment involves constructing a feature vector from the non-inductive load characteristic index and constructing an attribute vector from the pre-acquired electrical attributes of the power supply unit. This includes: calculating the normalized values ​​of the non-inductive load characteristic index based on its original values; constructing a feature vector based on the normalized values ​​of the non-inductive load characteristic index; acquiring the topology, electrical parameters, and power supply boundary of the power supply unit; and constructing an attribute vector based on the topology, electrical parameters, power supply boundary, and attribute dimension of the power supply unit.

[0042] One preferred implementation method, as shown in Figure 2, involves constructing a feature vector from the insensible load characteristic index of each load point, and highlighting key indicators through normalization and weight setting. The specific calculation formula is as follows: ; In the formula, Let i be the original value of the load point i at the kth inductive load characteristic index; Let j be the original value of the load point j at the kth inductive load characteristic index; For load point i, the normalized value of the kth inductive load characteristic index, such as peak-to-valley ratio, volatility, and regulation potential; For the set of load points; is the non-inductive load characteristic vector of load point i; m is the total number of non-inductive load characteristic indices, which is generally set to m=7.

[0043] Then, the electrical attributes of the power supply unit are converted into attribute vectors, and a normalized difference measure is used to eliminate the influence between different dimensions. The specific calculation formula is as follows: In the formula, This is the attribute vector of the nth power supply unit; Let p be the p-th electrical attribute of the nth power supply unit, such as node voltage, line impedance, power supply radius, and boundary constraints; p is the number of dimensions of the power supply unit attribute.

[0044] In the above scheme, the characteristics of non-inductive loads and the electrical attributes of power supply units are constructed as feature vectors and attribute vectors, respectively. The complex load and power grid entities are abstracted into vectors in mathematical space, which effectively eliminates the dominant influence of different dimensions on the calculation results, provides a reliable data foundation for the construction of subsequent evaluation models, and ensures that the zoning results are not only theoretically optimal, but also meet the safety and stability requirements of the power grid in practical applications, thereby improving the reliability of power grid operation.

[0045] A preferred approach involves constructing a two-dimensional evaluation model based on feature vectors and attribute vectors, including: calculating a measure of the difference in non-inductive load characteristics based on feature vectors, a first weight of the non-inductive load characteristic index, and a preset cluster center; calculating a measure of the difference in the electrical attributes of the power supply unit based on attribute vectors, a second weight of the electrical attributes of the power supply unit, and a preset cluster center; summing the measure of the difference in non-inductive load characteristics and the measure of the difference in the electrical attributes of the power supply unit, and setting weight coefficients in the corresponding dimensions of the measure of the difference in non-inductive load characteristics and the measure of the difference in the electrical attributes of the power supply unit to obtain the two-dimensional evaluation model.

[0046] One preferred implementation method, as shown in Figure 2, involves calculating the difference in non-inductive load characteristics using feature vectors. The calculation formula is as follows: In the formula, Let load point i and cluster center be Differences in the dimension of non-inductive load characteristics; Cluster center The mean value over the kth inductive load characteristic; The weight of the k-th inductive load characteristic index is represented by the first weight of the inductive load characteristic index, satisfying... In this embodiment, the first weight of the non-intrusive load characteristic index can be achieved using expert experience methods such as the Analytic Hierarchy Process (AHP) and objective statistical methods such as the entropy weight method, or a combination of both can be used to balance engineering experience and data differences. When prior information is lacking, each characteristic can also be assigned an equal weight. In the analysis of residential and commercial load characteristics, volatility, peak-to-valley ratio, and regulation potential have a greater impact on the zoning results and are often given higher weights. Due to the strong continuity of industrial loads, load factors and maximum ramp rate are more meaningful for distinguishing differences, and their weights can be appropriately increased. New infrastructure loads exhibit high stability during operation, and their characteristic differences are mainly reflected in complementarity and correlation.

[0047] The electrical attribute difference measure of the power supply unit is calculated using attribute vectors. The specific calculation formula is as follows: In the formula, For power supply unit n and cluster center Differences in power supply unit attributes; This refers to the q-th electrical attribute of the n-th power supply unit. Cluster center The mean value of the attribute of the qth power supply unit; and The maximum and minimum values ​​of this attribute in the sample are used for normalization; The weight of the q-th power supply unit attribute is represented by the second weight of the electrical attributes of the power supply unit, satisfying the following condition. In this embodiment, the second weight of the electrical attributes of the power supply unit can be determined by combining the actual needs of power grid planning and the optimization objectives of the zoning. Specifically, it can be obtained in the following ways: First, according to the priority setting of power grid planning standards, the power supply radius has a higher weight, generally 0.25-0.30, the reserve capacity weight is 0.15-0.20, and the weight of the number of transferable units and the N-1 verification result is 0.10-0.15; Second, it can be determined by using the entropy weight method or variance analysis based on historical cases and operating data; Third, the initial value is corrected by combining expert review and simulation sensitivity analysis.

[0048] Finally, a two-dimensional evaluation model of inductive load characteristics and power supply unit electrical attribute differences is constructed based on the inductive load characteristic difference measure and the power supply unit electrical attribute difference measure. The specific expression is as follows: In the formula, This is the comprehensive evaluation value of load point i relative to power supply unit n; The degree of difference in non-inductive load characteristics; For the degree of difference in power supply unit attributes; Let i be the vector of the inductive load characteristics at load point i; This is the attribute vector of the nth power supply unit; As a cluster center in the joint feature space, it is used to simultaneously characterize the relationship with non-inductive load characteristics and the relationship with the electrical attributes of the power supply unit, so that the difference between the two dimensions can be comprehensively evaluated under a unified spatial reference. and The weight coefficients for the two dimensions satisfy... In this embodiment, The zoning optimization objectives can be set accordingly. Specifically, when the zoning is mainly used for sensorless load regulation and demand response research, the following can be taken: When the zoning is mainly used for engineering feasibility and power supply security analysis, the appropriate zoning method can be adopted. In comprehensive scenarios, it is advisable to... In addition, a data-driven approach can be used to determine the weights, specifically by automatically determining the weights based on the variance contribution rates of the two types of differences, or by correcting them during simulation iterations through sensitivity analysis.

[0049] In the above scheme, the difference in characteristics of inductive loads and the difference in electrical attributes of power supply units are calculated using feature vectors and attribute vectors, respectively. Adjustable weight coefficients are introduced in the calculation process to establish a two-dimensional evaluation model that considers the similarity and complementarity of load characteristics within the region and the electrical carrying capacity and boundaries of the power supply unit itself. The two-dimensional evaluation model outputs a zoning scheme that meets the needs of specific scenarios, effectively preventing the problems of load concentration and capacity overrun. It can ensure that the zoning results are not only theoretically optimal, but also meet the safety and stability requirements of the power grid in practical applications, thereby improving the reliability of power grid operation.

[0050] A preferred scheme involves obtaining several candidate zones based on pre-acquired electrical attributes of power supply units, calculating the comprehensive evaluation value of each load point and each power supply unit using a two-dimensional evaluation model, and allocating each load point to a candidate zone that meets the preset zoning requirements based on the comprehensive evaluation value, thereby obtaining a preliminary zoning scheme. This includes: performing a preliminary combination of each load point and its corresponding power supply unit based on the pre-acquired electrical attributes of the power supply units to obtain several candidate zones; calculating the difference in inductive load characteristics and the difference in electrical attributes of each power supply unit between each load point and each power supply unit using a two-dimensional evaluation model to obtain the comprehensive evaluation value of each load point and each power supply unit; constructing a comprehensive evaluation matrix based on the comprehensive evaluation value of each load point and each power supply unit; and allocating each load point to a candidate zone that meets the preset zoning requirements based on the comprehensive evaluation matrix to obtain the preliminary zoning scheme.

[0051] As shown in Figure 2, one preferred implementation method involves initially combining load points and their power supply units based on the physical boundaries and electrical topology of the power supply units, taking into account factors such as power supply radius, line connection relationships, and substation service area. This forms several candidate partitions that meet basic geographical and electrical constraints. Each candidate partition is a collection of multiple power supply units and serves as the basic analysis unit for subsequent optimization. Then, a dual-dimensional evaluation model of non-inductive load characteristics and power supply units is used to calculate the matching degree. In this embodiment, the matching degree is represented by the difference in non-inductive load characteristics between each load point and each power supply unit, and the difference in electrical attributes between the power supply units. At the power supply unit level, for each load point, the difference between it and the cluster centers of each power supply unit is calculated in both the non-inductive load characteristic dimension and the power supply unit attribute dimension to obtain the comprehensive evaluation value of the load point-power supply unit. At the candidate partition level, when multiple power supply units are combined to form a candidate partition, the comprehensive evaluation values ​​of each power supply unit within the partition are weighted and summarized to obtain the comprehensive evaluation matrix of the load point-candidate partition. The calculation formula is as follows: ; Candidate partition Composed of several power supply units When assembling, define load point i as a candidate partition. The comprehensive evaluation matrix; where, Power supply unit n in candidate partition The weights within can be set according to capacity margin, node importance, or equal weight; Let be the comprehensive evaluation value of load point i relative to power supply unit n. It is worth noting that the comprehensive evaluation matrix of load point-candidate partitions reflects the rationality of assigning each load point to different candidate partitions, judges the rationality of local partitioning, and provides a quantitative basis for subsequent local partitioning and global optimization. Finally, based on the comprehensive evaluation matrix of load point-candidate partitions, load points are preferentially assigned to the candidate partitions with the highest matching degree, resulting in a preliminary partitioning scheme, i.e., the preliminary partitioning scheme. The specific expression of the preliminary partitioning assignment rule is as follows: In the formula, This is the partition index to which load point i belongs in the initial partitioning scheme. Candidate partition Composed of several power supply units When assembling, define load point i as a candidate partition. The comprehensive evaluation matrix.

[0052] Set of preliminary partitioning schemes The expression is: In the formula, H represents the number of candidate partitions. In the above scheme, the electrical attributes of the power supply units are used to initially combine each load point and its corresponding power supply unit, generating several candidate partitions. Then, a two-dimensional evaluation model is used to calculate the difference in inductive load characteristics and the difference in electrical attributes of each power supply unit between each load point and each power supply unit. This yields a comprehensive evaluation value for each load point and each power supply unit. A comprehensive evaluation matrix is ​​then constructed to allocate each load point to a candidate partition that meets the preset partitioning requirements, resulting in a preliminary partitioning scheme. Thus, each load point is assigned to the region that best matches its characteristics and is most compatible with the attributes of the power supply unit under the given conditions. This ensures that the generated preliminary partitioning scheme has reached a relatively optimal state in a localized area, effectively preventing load concentration and capacity overruns. It ensures that the partitioning results are not only theoretically optimal but also meet the safety and stability requirements of the power grid in practical applications, improving the reliability of power grid operation.

[0053] A preferred approach involves constructing a global optimization model based on a preliminary partitioning scheme, optimizing each candidate partition, and obtaining a preliminary partitioning optimization scheme. This includes: constructing a global optimization model based on the preliminary partitioning scheme and a preset objective function; optimizing the attribution relationship between each load point and each candidate partition until the preset global partitioning optimization requirements are met, thus obtaining a preliminary partitioning optimization scheme.

[0054] As shown in Figure 2, one possible implementation of a preferred scheme involves obtaining a preliminary partitioning scheme. While the preliminary partitioning scheme has high rationality within a local range, it may still have issues such as uneven load distribution between partitions and violation of local capacity constraints. Therefore, further global optimization is required. This involves establishing an objective function based on the preliminary partitioning as input, resulting in a global optimization model. The global optimization model optimizes in two main ways: firstly, minimizing the comprehensive evaluation value between load points within a partition and the partition itself, i.e., the weighted sum of the comprehensive evaluation matrix; secondly, maximizing the complementarity of different loads within a partition in terms of time-series characteristics. Simultaneously, to ensure the engineering feasibility and operational safety of the results, a power supply radius constraint and an "N-1" safety check constraint are introduced during the optimization process. Under the dual influence of the evaluation matrix and the global optimization objective function, the load point attribution relationship is adjusted multiple times, continuously eliminating unreasonable partitions and improving partition balance until convergence to a globally optimal or suboptimal partitioning scheme, thus obtaining the preliminary partitioning optimization scheme. In this embodiment, the preset global partitioning optimization requirement is characterized by the partitioning scheme converging to a globally optimal or suboptimal state.

[0055] The objective function expression of the global optimization model is: In the formula, Z is the global optimization objective value; Let h be the set of load points contained in the h-th partition; The complementarity index between load point i and load point j; Candidate partition Composed of several power supply units When assembling, define load point i as a candidate partition. The comprehensive evaluation matrix is ​​then used to further determine the power supply radius constraint, assuming the distribution network topology is... It means that, among them: For a set of nodes, For the collection of routes, The weight of the edge is the electrical impedance. Therefore, the distance from load point i to the cluster center of partition h is... gather electrical shortest path length It can be defined as: ; In the formula, From load point i to cluster center The set of all feasible paths; This is the weighted sum of all edges on path p; The weight of the edge; Let h be the set of load points contained in the h-th partition; The maximum power supply radius set for the system.

[0056] The "N-1" safety check constraint is further defined, and its specific expression is as follows: In the formula, The total load within partition h; This is the maximum safe load of the main power supply line for this zone; The maximum allowable load rate is typically set between 0.5 and 0.8.

[0057] In the above scheme, a global optimization model is constructed from the preliminary zoning scheme and a preset objective function. The relationship between each load point and each candidate zone is optimized, and the load balance between zones and the deep complementarity of loads within the region, which were ignored in the preliminary zoning scheme, are optimized. This effectively prevents the problems of load concentration and capacity overrun. It ensures that the zoning results are not only theoretically optimal, but also meet the safety and stability requirements of the power grid in practical applications, thereby improving the reliability of power grid operation.

[0058] A preferred scheme involves constructing corresponding power flow balance equations based on a preliminary zoning optimization scheme, and locally adjusting the preliminary zoning optimization scheme to obtain a distribution network zoning optimization scheme for optimizing the distribution network zoning. This includes: performing preliminary zoning of the distribution network based on the preliminary zoning optimization scheme to obtain zoning results; constructing power flow balance equations corresponding to each zoning result, and mapping load points to corresponding power flow nodes according to the power flow balance equations; performing power flow verification on each power flow node; if the power flow verification results do not meet preset operating requirements, locally adjusting the preliminary zoning optimization scheme until the power flow verification results meet the preset operating requirements; and constructing an operating optimization model to perform operating optimization on the preliminary zoning optimization scheme that meets the preset operating requirements to obtain the distribution network zoning optimization scheme for optimizing the distribution network zoning.

[0059] One preferred implementation method is shown in Figure 2. After completing the partition optimization, the partition results are obtained according to the preliminary partition optimization scheme and the corresponding power flow balance equation is established. The voltage of each node, the line current and the system power balance are checked. The partition results are evaluated for voltage deviation, line overload and network loss. If the operation results show that there are violations of operation constraints such as voltage exceeding the limit or line overload, some load points are locally adjusted or redistributed. The specific execution steps are as follows: the preliminary partition optimization scheme is returned to the stage of step 104 to step 105 for processing, and the infeasible preliminary partition optimization scheme is masked. The partition optimization is re-executed. This constructs a closed-loop iterative process for zoning optimization and power flow verification until a zoning scheme that meets operational safety requirements is obtained. In this embodiment, the preset operational requirements are characterized by the absence of operational constraint violations such as voltage overruns or line overloads. In this embodiment, during the construction of the power flow balance equation, there is a correspondence between load points and nodes in the power flow calculation. That is, a load point refers to a user or load unit with independent power consumption characteristics in the distribution network, whose power demand is connected to the grid through distribution lines; a node is the electrical bus in the power flow analysis, representing the connection point of the electrical network. To facilitate unified modeling, each load point is mapped to the electrical node it connects to, with the following specific rules: when a node connects to only a single load point, there is a one-to-one correspondence between the load point and the node; when a node connects to multiple load points, the total load of the node can be defined as the power superposition of all connected load points; when a node connects to multiple load points, the total load of the node can be defined as the power superposition of all connected load points; through the above mapping relationship, the load curve data of the load point is obtained. It can be directly used as the input for node injection power, thus ensuring the consistency of the symbol system between the partition optimization and power flow verification models. The specific expression of the power flow balance equation is as follows: ; ; In the formula, and For nodes The active and reactive power injected; This represents the total number of system nodes. and For nodes and The voltage amplitude; The phase angle difference of the node voltage; and For nodes and The real and imaginary parts of the admittance; and These are the allowable upper and lower limits for node voltage. For nodes The voltage amplitude.

[0060] After the partitioning passes the operational security verification, an operational optimization model is constructed. By minimizing the system network loss level and partition load imbalance, the economic efficiency and overall efficiency of system operation can be effectively improved. The specific expression of the operational optimization objective function of the operational optimization model is as follows: In the formula, F is the objective function for runtime optimization; and These are weighting coefficients, generally , .

[0061] in, The system network loss is calculated using the following formula: In the formula, For the line - The resistance; For line current constraints, the expression for line current constraints is: In the formula, For the line - Current flow; For the line - The maximum permissible current.

[0062] Based on the verification results, the preliminary zoning optimization scheme is partially adjusted, and the final distribution network zoning optimization scheme is generated to achieve the optimization of the distribution network zoning.

[0063] In the above scheme, the partitioning results are obtained from the preliminary partitioning optimization scheme, the power flow balance equations corresponding to each partitioning result are constructed, and the load points are mapped to the corresponding power flow nodes. Finally, the preliminary partitioning optimization scheme is optimized in operation, and then the preliminary partitioning optimization scheme is further adjusted locally. The introduction of a power flow verification mechanism effectively prevents the problems of load concentration and capacity over-limit. It can ensure that the partitioning results are not only theoretically optimal, but also meet the safety and stability requirements of the power grid in practical applications, thereby improving the reliability of power grid operation.

[0064] Example 2 refers to Figure 3, which is a schematic diagram of the module structure of a distribution network zoning optimization system provided in a certain embodiment of the present invention. As shown in Figure 3, the embodiment of the present invention also provides a distribution network zoning optimization system, including: a characteristic index set construction module 201, a vector construction module 202, a two-dimensional evaluation model construction module 203, a preliminary zoning scheme acquisition module 204, a preliminary zoning optimization scheme acquisition module 205, and a distribution network zoning optimization module 206; the characteristic index set construction module 201 is used to construct a set of non-inductive load characteristic indices based on pre-acquired original load point data; the vector construction module 202 is used to construct non-inductive load characteristic indices into feature vectors and construct pre-acquired power supply unit electrical attributes into attribute vectors; the two-dimensional evaluation model construction module 203 is used to construct a two-dimensional evaluation model based on the feature vectors and attribute vectors. The system comprises the following modules: a preliminary zoning scheme acquisition module 204, which acquires several candidate zoning schemes based on the pre-acquired electrical attributes of power supply units, calculates the comprehensive evaluation value of each load point and each power supply unit through a two-dimensional evaluation model, and allocates each load point to a candidate zoning scheme that meets the preset zoning requirements based on the comprehensive evaluation value, thereby obtaining a preliminary zoning scheme; a preliminary zoning optimization scheme acquisition module 205, which constructs a global optimization model based on the preliminary zoning scheme, optimizes each candidate zoning scheme, and obtains a preliminary zoning optimization scheme; and a distribution network zoning optimization module 206, which constructs the corresponding power flow balance equation based on the preliminary zoning optimization scheme, and locally adjusts the preliminary zoning optimization scheme to obtain a distribution network zoning optimization scheme for optimizing the distribution network zoning.

[0065] This invention proposes a distribution network zoning optimization system. It extracts non-inductive load characteristic indicators from raw load point data, constructs these indicators as feature vectors, and then constructs attribute vectors from pre-acquired electrical attributes of power supply units. This fully considers the flexibility and adjustability of the loads, as well as the rigidity and structural parameters of the power grid, providing a reliable data foundation for subsequent optimization analysis. A two-dimensional evaluation model is then constructed based on the feature vectors and attribute vectors to comprehensively evaluate each load point and each power supply unit, generating a preliminary zoning scheme. This preliminary zoning scheme is then globally optimized to obtain a preliminary optimized zoning scheme, effectively preventing load concentration and capacity overruns. Finally, a power flow balance equation is constructed to locally adjust the preliminary optimized zoning scheme to generate a final optimized distribution network zoning scheme. Therefore, by introducing a two-dimensional evaluation model and combining non-inductive load characteristics with the electrical attributes of power supply units, the system effectively improves the load balance of the distribution network and the overall engineering feasibility of the scheme. This ensures that the zoning results are not only theoretically optimal but also meet the safety and stability requirements of the power grid in practical applications, thereby improving the reliability of power grid operation.

[0066] Furthermore, the characteristic index set construction module 201 is used to construct a set of non-intrusive load characteristic indicators based on the pre-acquired raw load point data, including: a first data acquisition unit 301, a matrix construction unit 302, and a characteristic index extraction unit 303; the data acquisition unit 301 is used to collect the spatial coordinates of the load points, the predicted annual load data, and the typical daily load curves to obtain the raw load point data; the matrix construction unit 302 is used to construct the raw daily load curve matrix based on the raw load point data; the characteristic index extraction unit 303 is used to preprocess the raw daily load curve matrix and extract several characteristic indicators from the raw daily load curve matrix to obtain the set of non-intrusive load characteristic indicators.

[0067] Furthermore, the vector construction module 202 is used to construct the non-inductive load characteristic index into a feature vector and the pre-acquired power supply unit electrical attributes into an attribute vector, including: a feature vector construction unit 401, a second data acquisition unit 402, and an attribute vector construction unit 403; the feature vector construction unit 401 is used to calculate the normalized value of the non-inductive load characteristic index based on the original value of the non-inductive load characteristic index, and construct a feature vector based on the normalized value of the non-inductive load characteristic index; the second data acquisition unit 402 is used to acquire the topology, electrical parameters, and power supply boundary of the power supply unit; the attribute vector construction unit 403 is used to construct an attribute vector based on the topology, electrical parameters, power supply boundary, and attribute dimension of the power supply unit.

[0068] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

[0069] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. Furthermore, the described specific features, structures, materials, or characteristics may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of those different embodiments or examples.

[0070] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "a plurality of" means two or more, unless otherwise explicitly specified.

Claims

1. A method for optimizing the zoning of a power distribution network, characterized in that, include: Based on the pre-acquired raw load point data, a set of non-inductive load characteristic indicators is constructed; the non-inductive load characteristic indicators are constructed as feature vectors, and the pre-acquired electrical attributes of power supply units are constructed as attribute vectors; a two-dimensional evaluation model is constructed based on the feature vectors and the attribute vectors; several candidate partitions are obtained based on the pre-acquired electrical attributes of power supply units, and the comprehensive evaluation value of each load point and each power supply unit is calculated through the two-dimensional evaluation model, and each load point is assigned to a candidate partition that meets the preset partitioning requirements based on the comprehensive evaluation value, thus obtaining a preliminary partitioning scheme; Based on the preliminary partitioning scheme, a global optimization model is constructed, and each candidate partition is optimized to obtain a preliminary partitioning optimization scheme. Based on the preliminary zoning optimization scheme, the corresponding power flow balance equation is constructed, and the preliminary zoning optimization scheme is locally adjusted to obtain the distribution network zoning optimization scheme, so as to optimize the distribution network zoning.

2. The distribution network zoning optimization method as described in claim 1, characterized in that, The process of constructing a set of non-intrusive load characteristic indicators based on pre-acquired raw load point data includes: collecting the spatial coordinates of load points, predicted annual load data, and typical daily load curves to obtain raw load point data; constructing a raw daily load curve matrix based on the raw load point data; preprocessing the raw daily load curve matrix and extracting several characteristic indicators from the raw daily load curve matrix to obtain a set of non-intrusive load characteristic indicators.

3. The distribution network zoning optimization method as described in claim 1, characterized in that, The non-inductive load characteristic index is constructed into a feature vector, and the pre-acquired electrical attributes of the power supply unit are constructed into an attribute vector, including: calculating the normalized value of the non-inductive load characteristic index based on the original value of the non-inductive load characteristic index, constructing a feature vector based on the normalized value of the non-inductive load characteristic index; acquiring the topology, electrical parameters, and power supply boundary of the power supply unit; and constructing an attribute vector based on the topology, electrical parameters, power supply boundary, and attribute dimension of the power supply unit.

4. The distribution network zoning optimization method as described in claim 1, characterized in that, A two-dimensional evaluation model is constructed based on the feature vector and the attribute vector, including: calculating the difference measure of non-inductive load characteristics based on the feature vector, the first weight of the non-inductive load characteristic index, and the preset cluster center; calculating the difference measure of power supply unit electrical attributes based on the attribute vector, the second weight of the power supply unit electrical attributes, and the preset cluster center; summing the difference measure of non-inductive load characteristics and the difference measure of power supply unit electrical attributes, and setting weight coefficients in the corresponding dimensions of the difference measure of non-inductive load characteristics and the difference measure of power supply unit electrical attributes to obtain the two-dimensional evaluation model.

5. The distribution network zoning optimization method as described in claim 4, characterized in that, Based on the pre-acquired electrical attributes of the power supply units, several candidate partitions are obtained. The comprehensive evaluation value of each load point and each power supply unit is calculated using the dual-dimensional evaluation model. The load points are then assigned to candidate partitions that meet the preset partitioning requirements based on the comprehensive evaluation value, resulting in a preliminary partitioning scheme. This includes: performing a preliminary combination of each load point and its corresponding power supply unit based on the pre-acquired electrical attributes of the power supply units to obtain several candidate partitions; calculating the difference in inductive load characteristics and the difference in electrical attributes of each power supply unit between each load point and each power supply unit based on the dual-dimensional evaluation model to obtain a comprehensive evaluation value for each load point and each power supply unit; constructing a comprehensive evaluation matrix based on the comprehensive evaluation values ​​of each load point and each power supply unit; and assigning each load point to candidate partitions that meet the preset partitioning requirements based on the comprehensive evaluation matrix, thus obtaining a preliminary partitioning scheme.

6. The distribution network zoning optimization method as described in claim 1, characterized in that, Based on the preliminary partitioning scheme, a global optimization model is constructed, and each candidate partition is optimized to obtain a preliminary partitioning optimization scheme. This includes: constructing a global optimization model based on the preliminary partitioning scheme and a preset objective function; optimizing the attribution relationship between each load point and each candidate partition until the preset global partitioning optimization requirements are met, thus obtaining a preliminary partitioning optimization scheme.

7. The distribution network zoning optimization method as described in claim 6, characterized in that, Based on the preliminary zoning optimization scheme, a corresponding power flow balance equation is constructed, and the preliminary zoning optimization scheme is locally adjusted to obtain a distribution network zoning optimization scheme for optimizing the distribution network zoning. This includes: performing preliminary zoning of the distribution network based on the preliminary zoning optimization scheme to obtain zoning results; constructing power flow balance equations corresponding to each zoning result, and mapping the load points to corresponding power flow nodes according to the power flow balance equations; performing power flow verification on each power flow node; if the power flow verification result does not meet the preset operating requirements, then locally adjusting the preliminary zoning optimization scheme until the power flow verification result meets the preset operating requirements, and constructing an operating optimization model to perform operating optimization on the preliminary zoning optimization scheme that meets the preset operating requirements to obtain a distribution network zoning optimization scheme for optimizing the distribution network zoning.

8. A distribution network zoning optimization system, characterized in that, The distribution network zoning optimization method according to any one of claims 1 to 7 includes: a characteristic index set construction module, a vector construction module, a two-dimensional evaluation model construction module, a preliminary zoning scheme acquisition module, a preliminary zoning optimization scheme acquisition module, and a distribution network zoning optimization module; the characteristic index set construction module is used to construct an inductive load characteristic index set based on pre-acquired original load point data; the vector construction module is used to construct the inductive load characteristic index as a feature vector and construct the pre-acquired power supply unit electrical attributes as an attribute vector; the two-dimensional evaluation model construction module is used to construct a two-dimensional evaluation model based on the feature vector and the attribute vector; the preliminary zoning scheme acquisition module... The module is used to obtain several candidate partitions based on the pre-acquired electrical attributes of the power supply units, calculate the comprehensive evaluation value of each load point and each power supply unit through the dual-dimensional evaluation model, and allocate each load point to a candidate partition that meets the preset partitioning requirements based on the comprehensive evaluation value to obtain a preliminary partitioning scheme; the preliminary partitioning optimization scheme acquisition module is used to construct a global optimization model based on the preliminary partitioning scheme, optimize each candidate partition, and obtain a preliminary partitioning optimization scheme; the distribution network partitioning optimization module is used to construct the corresponding power flow balance equation based on the preliminary partitioning optimization scheme, and locally adjust the preliminary partitioning optimization scheme to obtain a distribution network partitioning optimization scheme to optimize the distribution network partitioning.

9. A distribution network zoning optimization system as described in claim 8, characterized in that, The characteristic index set construction module is used to construct a set of non-impact load characteristic indicators based on pre-acquired raw load point data, including: a first data acquisition unit, a matrix construction unit, and a characteristic index extraction unit; the data acquisition unit is used to collect the spatial coordinates of load points, predicted annual load data, and typical daily load curves to obtain raw load point data; the matrix construction unit is used to construct a raw daily load curve matrix based on the raw load point data; the characteristic index extraction unit is used to preprocess the raw daily load curve matrix and extract several characteristic indicators from the raw daily load curve matrix to obtain a set of non-impact load characteristic indicators.

10. A distribution network zoning optimization system as described in claim 8, characterized in that, The vector construction module is used to construct the non-inductive load characteristic index into a feature vector and to construct the pre-acquired electrical attributes of the power supply unit into an attribute vector. It includes a feature vector construction unit, a second data acquisition unit, and an attribute vector construction unit. The feature vector construction unit is used to calculate the normalized value of the non-inductive load characteristic index based on its original value, and to construct a feature vector based on the normalized value of the non-inductive load characteristic index. The second data acquisition unit is used to acquire the topology, electrical parameters, and power supply boundary of the power supply unit. The attribute vector construction unit is used to construct an attribute vector based on the topology, electrical parameters, power supply boundary, and attribute dimension of the power supply unit.

Images