A power distribution network transformer capacity configuration and collaborative control method and system

Abstract

The application relates to the technical field of power distribution networks, and discloses a power distribution network transformer capacity configuration and collaborative control method and system. The method comprises the following steps: obtaining historical load data of a target area in a power distribution network, decomposing and predicting the historical load data, and determining a load density variation trend of the target area; performing semantic analysis on the load density variation trend, performing multi-step modeling on obtained load semantic features, and obtaining a first transformer capacity configuration strategy of the target area; obtaining real-time load data of the target area, performing matching degree calculation on the real-time load data and the first transformer capacity configuration strategy, optimizing the first transformer capacity configuration strategy based on the obtained matching degree, and obtaining a second transformer capacity configuration strategy of the target area; and obtaining a transformer collaborative control strategy of the target area based on real-time operation data by using a particle swarm optimization algorithm. The application improves the operation efficiency and reliability of the power distribution network.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power distribution network, in particular to a power distribution network transformer capacity configuration and collaborative control method and system. BACKGROUND

[0002] In the power distribution network, the load density of the transformer area is an important factor affecting the selection of the transformer, and its change trend directly determines the capacity demand of the transformer. The load density of the transformer area may show a trend of gradual increase or fluctuation, which puts higher requirements on the initial configuration and later operation of the transformer capacity. During the operation of the transformer, the load curve characteristics (such as peak-valley difference and load fluctuation frequency) reflect the change law of the actual electricity demand. The matching degree between these characteristics and the transformer capacity margin directly affects the operation efficiency and service life of the transformer. If the capacity margin is too large, the transformer is in a low load state for a long time, resulting in low resource utilization; if the capacity margin is too small, the transformer may be frequently overloaded, increasing the equipment loss and failure risk. At present, there is a lack of effective linkage mechanism among multiple transformers, resulting in uneven load distribution, individual transformer long-term overload and other transformers low-load operation, affecting the efficiency and reliability of the entire power distribution network.

[0003] Therefore, it is necessary to establish an intelligent linkage mechanism for the transformer group to realize the dynamic load distribution and capacity optimization among the transformers, so as to ensure the efficiency and reliability of the power distribution network. SUMMARY

[0004] In view of the above problems, the present application provides a power distribution network transformer capacity configuration and collaborative control method and system.

[0005] In a first aspect, the embodiments of the present application provide a power distribution network transformer capacity configuration and collaborative control method, comprising:

[0006] Obtaining historical load data of a target transformer area in a power distribution network, and decomposing and predicting the historical load data to determine the load density change trend of the target transformer area;

[0007] Performing semantic analysis on the load density change trend, and performing multi-step modeling on the obtained load semantic features to obtain a first transformer capacity configuration strategy of the target transformer area, wherein the multi-step modeling includes intra-day load feature matrix modeling, transformer capacity prediction modeling and transformer capacity configuration modeling;

[0008] Obtaining real-time load data of the target transformer area, performing matching degree calculation on the real-time load data and the first transformer capacity configuration strategy, and optimizing the first transformer capacity configuration strategy based on the obtained matching degree to obtain a second transformer capacity configuration strategy of the target transformer area;

[0009] acquire real-time operation data of each transformer in the target transformer area, and obtain a transformer cooperative control strategy of the target transformer area based on the real-time operation data by using a particle swarm optimization algorithm, wherein the transformer cooperative control strategy comprises an active power regulation target value, a reactive power regulation target value and a voltage regulation target value of each transformer.

[0010] Preferably, the historical load data is decomposed and predicted to determine the load density variation trend of the target transformer area, comprising:

[0011] The historical load data is decomposed and predicted by using a discrete wavelet transform to obtain a multi-time granularity load data sequence;

[0012] The multi-time granularity load data sequence is extracted by using a first recurrent neural network, and at least based on the obtained time feature vector, load density prediction data of the target transformer area is obtained by performing load density prediction;

[0013] The load density prediction data is clustered, and a load density growth trend of the target transformer area is identified based on the clustering result.

[0014] Preferably, the load density variation trend is semantically analyzed, and the obtained load semantic features are modeled in multiple steps to obtain a first transformer capacity configuration strategy of the target transformer area, comprising:

[0015] The load density variation trend is semantically analyzed by using a large language model to obtain load semantic features;

[0016] Based on the load semantic features, the historical load data is subjected to ultra-limit statistics, and an intra-day load feature matrix of the target transformer area is constructed according to the statistical result, wherein the intra-day load feature matrix comprises load features at a plurality of time points in a day, and the load features comprise peak load, duration and occurrence period;

[0017] The intra-day load feature matrix is subjected to capacity mapping by using a second recurrent neural network, and a plurality of transformer capacity levels are obtained by clustering the obtained transformer capacity prediction data;

[0018] Each transformer capacity level is prioritized based on an analytic hierarchy process, and the transformer capacity level with the highest priority is represented as the first transformer capacity configuration strategy of the target transformer area.

[0019] Preferably, the real-time load data and the first transformer capacity configuration strategy are matched to calculate a matching degree, and the first transformer capacity configuration strategy is optimized based on the obtained matching degree to obtain a second transformer capacity configuration strategy of the target transformer area, comprising:

[0020] statistically over-limiting the real-time load data based on a transformer load rate threshold, and normalizing the statistical result to obtain a transformer capacity matching quantization value;

[0021] calculating a transformer capacity margin coefficient corresponding to the first transformer capacity configuration strategy based on a support vector regression;

[0022] performing weighted summation on the transformer capacity matching quantization value and the transformer capacity margin coefficient to obtain a matching degree of the real-time load data and the first transformer capacity configuration strategy;

[0023] judging whether the matching degree is higher than a first preset threshold, and if not, iteratively optimizing the first transformer capacity configuration strategy to obtain a second transformer capacity configuration strategy of the target transformer area.

[0024] Preferably, the iteratively optimizing the first transformer capacity configuration strategy to obtain the second transformer capacity configuration strategy of the target transformer area comprises:

[0025] iteratively optimizing the first transformer capacity configuration strategy by using a gradient descent method until a first convergence condition is reached to obtain the second transformer capacity configuration strategy of the target transformer area, wherein the iteratively optimizing comprises adjusting a transformer capacity level and a transformer capacity margin coefficient, and the first convergence condition is that a difference in the matching degree of two consecutive iterations is less than a second preset threshold.

[0026] Preferably, the obtaining the transformer collaborative control strategy of the target transformer area based on the real-time operation data by using a particle swarm optimization algorithm comprises:

[0027] obtaining a first transformer collaborative control strategy of the target transformer area based on the real-time operation data, wherein the first transformer collaborative control strategy comprises real-time active power, real-time reactive power and real-time voltage of each transformer;

[0028] iteratively optimizing the first transformer collaborative control strategy by using a particle swarm optimization algorithm until a second convergence condition is reached to obtain a second transformer collaborative control strategy of the target transformer area, wherein the iteratively optimizing comprises calculating a load imbalance degree of the target transformer area based on a load rate of each transformer, and the second convergence condition is that a difference in the load imbalance degree of three consecutive iterations is less than a third preset threshold.

[0029] characterizing the second transformer collaborative control strategy as the transformer collaborative control strategy of the target transformer area.

[0030] In a second aspect, an embodiment of the present application provides a power distribution network transformer capacity configuration and collaborative control system, comprising:

[0031] a decomposition prediction module, configured to acquire historical load data of a target transformer area in a power distribution network, and perform decomposition prediction on the historical load data to determine a load density variation trend of the target transformer area;

[0032] a first capacity configuration strategy determination module, configured to perform semantic analysis on the load density variation trend, and perform multi-step modeling on obtained load semantic features to obtain a first transformer capacity configuration strategy of the target transformer area, wherein the multi-step modeling comprises intra-day load feature matrix modeling, transformer capacity prediction modeling, and transformer capacity configuration modeling;

[0033] a second capacity configuration strategy determination module, configured to acquire real-time load data of the target transformer area, perform matching degree calculation on the real-time load data and the first transformer capacity configuration strategy, and perform optimization on the first transformer capacity configuration strategy based on the obtained matching degree to obtain a second transformer capacity configuration strategy of the target transformer area;

[0034] a collaborative control strategy determination module, configured to acquire real-time operation data of each transformer in the target transformer area, and obtain a transformer collaborative control strategy of the target transformer area based on the real-time operation data by using a particle swarm optimization algorithm, wherein the transformer collaborative control strategy comprises an active power regulation target value, a reactive power regulation target value, and a voltage regulation target value of each transformer.

[0035] Preferably, the decomposition prediction module comprises:

[0036] a multi-scale decomposition unit, configured to perform multi-scale decomposition on the historical load data by using discrete wavelet transform to obtain a multi-time granularity load data sequence;

[0037] a load density prediction unit, configured to perform data extraction on the multi-time granularity load data sequence by using a first recurrent neural network, and perform load density prediction based on at least a time feature vector obtained to obtain load density prediction data of the target transformer area;

[0038] a load density variation trend determination unit, configured to perform clustering on the load density prediction data, and identify a load density growth trend of the target transformer area based on a clustering result.

[0039] Preferably, the first capacity configuration strategy determination module comprises:

[0040] a semantic analysis unit, configured to perform semantic analysis on the load density variation trend by using a large language model to obtain load semantic features;

[0041] The day-to-day load feature matrix determination unit is configured to perform over-limit statistics on the historical load data based on the load semantic features, and construct a day-to-day load feature matrix of the target transformer area according to the statistical results, wherein the day-to-day load feature matrix comprises load features at a plurality of time points in a day, and the load features comprise peak load, duration and occurrence period;

[0042] The transformer capacity level determination unit is configured to perform capacity mapping on the day-to-day load feature matrix by using a second recurrent neural network, and perform clustering on the obtained transformer capacity prediction data to obtain a plurality of transformer capacity levels.

[0043] The priority ranking unit is configured to perform priority ranking on each of the transformer capacity levels based on an analytic hierarchy process, and represent the transformer capacity level with the highest priority as a first transformer capacity configuration strategy of the target transformer area.

[0044] Preferably, the second capacity configuration strategy determination module comprises:

[0045] The capacity matching degree quantization value determination unit is configured to perform over-limit statistics on the real-time load data based on a transformer load rate threshold, and perform normalization processing on the statistical results to obtain a transformer capacity matching degree quantization value.

[0046] The capacity margin coefficient determination unit is configured to calculate a transformer capacity margin coefficient corresponding to the first transformer capacity configuration strategy based on a support vector regression.

[0047] The matching degree calculation unit is configured to perform weighted summation on the transformer capacity matching degree quantization value and the transformer capacity margin coefficient to obtain a matching degree of the real-time load data and the first transformer capacity configuration strategy.

[0048] The iterative optimization unit is configured to determine whether the matching degree is higher than a first preset threshold, and if not, perform iterative optimization on the first transformer capacity configuration strategy to obtain a second transformer capacity configuration strategy of the target transformer area.

[0049] Compared with the prior art, the power distribution network transformer capacity configuration and collaborative control method and system of the embodiment of the present application has the beneficial effects that: by analyzing historical load data, the load density variation trend can be predicted, a first transformer capacity configuration strategy is generated by combining semantic analysis and multi-step modeling, the matching degree calculation of real-time load data and the first transformer capacity configuration strategy is introduced, the transformer capacity configuration can be optimized, for each transformer, a collaborative control strategy is formulated by using a particle swarm optimization algorithm, the intelligent configuration and group collaborative control of transformer capacity are realized, and the operation efficiency and reliability of the power distribution network are improved. BRIEF DESCRIPTION OF DRAWINGS

[0050] Figure 1 This is a flowchart illustrating a method for configuring and coordinating the capacity of a distribution network transformer according to an embodiment of the present invention.

[0051] Figure 2 This is a schematic diagram of the structure of a power distribution network transformer capacity configuration and coordinated control system according to an embodiment of the present invention;

[0052] Figure label:

[0053] 1. Decomposition and prediction module; 2. First capacity configuration strategy determination module; 3. Second capacity configuration strategy determination module; 4. Cooperative control strategy determination module. Detailed Implementation

[0054] The specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and are not intended to limit the scope of the invention.

[0055] In the description of this invention, it should be understood that the terms "first" and "second," etc., are used to distinguish different objects, rather than to describe a specific order.

[0056] In the description of this invention, it should be noted that, unless otherwise defined, all technical and scientific terms used in this invention have the same meaning as commonly understood by those skilled in the art. The terminology used in this specification is for the purpose of describing specific embodiments only and is not intended to limit the invention. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0057] like Figure 1 The diagram shown is a flowchart illustrating a method for distribution network transformer capacity configuration and coordinated control according to an embodiment of the present invention. (Refer to...) Figure 1 An embodiment of the present invention provides a method for distribution network transformer capacity configuration and coordinated control, comprising the following steps:

[0058] S1. Obtain historical load data of the target distribution area in the distribution network, and decompose and predict the historical load data to determine the load density change trend of the target distribution area.

[0059] Specifically, step S1 includes:

[0060] 1) Discrete wavelet transform is used to decompose historical load data into multi-scale data sequences with multiple time granularities.

[0061] The historical load data is decomposed into multi-scale by discrete wavelet transform, and the data granularity correspondence is established by different time scale coefficient mapping to obtain the multi-time granularity load data sequence. The historical load data in this embodiment is daily load data, which is converted into 15-minute granularity load data sequence by discrete wavelet transform.

[0062] 2) The first recurrent neural network is used to extract data from the multi-time granularity load data sequence, and at least based on the obtained time feature vector, the load density prediction data of the target area is obtained;

[0063] The first recurrent neural network is used to extract data from the multi-time granularity load data sequence, and at least based on the obtained time feature vector, the load density prediction data of the target area is obtained;

[0064] The first neural network in this embodiment is a long short-term memory network, which takes the time feature vector and the electricity peak coefficient as input when performing load density prediction, uses 128 neurons in the hidden layer, and outputs the load density prediction data of the target area. Among them, the electricity peak coefficient reflects the load peak-valley difference, and the larger the electricity peak coefficient, the more intense the load fluctuation.

[0065] 3) The load density prediction data is clustered, and the load density growth trend of the target area is identified based on the clustering result.

[0066] Specifically, the DBSCAN algorithm is used to cluster the load density prediction data, and the load density growth trend of the target area is identified based on the clustering result. The DBSCAN algorithm is a density-based spatial clustering algorithm, and the clustering radius is set to 0.15 and the minimum sample number is set to 4 in this embodiment. The load density growth trend can be effectively identified by the clustering result.

[0067] S2, the semantic analysis of the load density change trend is performed, and the obtained load semantic features are modeled in multiple steps to obtain the first transformer capacity configuration strategy of the target area;

[0068] Specifically, the multi-step modeling includes intra-day load feature matrix modeling, transformer capacity prediction modeling, and transformer capacity configuration modeling.

[0069] Further, step S2 includes:

[0070] 1) The semantic analysis of the load density change trend is performed by using a large language model to obtain load semantic features;

[0071] The DeepSeek large language model is used in this embodiment to perform semantic analysis on the load density change trend, and the "nighttime load is low, daytime fluctuates frequently, and double-peak characteristics appear in the afternoon and evening" load semantic characteristics are obtained.

[0072] 2) Perform over-limit statistics on historical load data based on the load semantic characteristics, and construct an intraday load feature matrix of the target substation according to the statistical results;

[0073] The load semantic characteristics are converted into numerical vectors and keywords are extracted using a text vector processing method. Based on the keywords, over-limit statistics are performed on historical load data, and an intraday load feature matrix of the target substation is constructed according to the statistical results. Specifically, the intraday load feature matrix includes load characteristics at several time points in a day, and the load characteristics include peak load, duration, and occurrence period.

[0074] In this embodiment, the load semantic characteristics are converted into numerical vectors and keywords are extracted, and the keywords include "frequent fluctuations", "double-peak characteristics", and "load concentration". The statistical results show that the target substation has a daily average duration of exceeding the rated capacity of 4 hours, and the maximum load rate reaches 120%. The intraday load feature matrix includes load characteristics at 24 time points in a day. That is, the size of the intraday load feature matrix is 24x3, corresponding to three dimensions of peak load, duration, and occurrence period.

[0075] 3) Use a second recurrent neural network to map the intraday load feature matrix to the transformer capacity, and perform clustering on the obtained transformer capacity prediction data to obtain several transformer capacity levels;

[0076] In this embodiment, the input layer node number of the second recurrent neural network is 24, corresponding to the load characteristics at each time point in a day, i.e., the intraday load feature matrix. The number of hidden layer nodes is 64, and the number of output layer nodes is 1, corresponding to the transformer capacity prediction data. The transformer capacity prediction data is clustered using k-means clustering, and three transformer capacity levels are obtained, which are small transformer capacity level 200 kVA, medium transformer capacity level 315 kVA, and high transformer capacity level 400 kVA.

[0077] 4) Prioritize each transformer capacity level based on the analytic hierarchy process, and represent the transformer capacity level with the highest priority as the first transformer capacity configuration strategy of the target substation.

[0078] In the construction of the judgment matrix by using the analytic hierarchy process, three evaluation criteria of economy, reliability and adaptability are set, and weights are respectively assigned. In the priority ranking of each transformer capacity level based on the analytic hierarchy process in this embodiment, the weights of the three evaluation criteria are 0.4, 0.35 and 0.25 respectively, and the transformer capacity level with the highest priority, i.e., the transformer capacity level of 315 kVA, is represented as the first transformer capacity configuration strategy of the target transformer area.

[0079] S3, obtaining real-time load data of the target transformer area, performing matching degree calculation on the real-time load data and the first transformer capacity configuration strategy, and optimizing the first transformer capacity configuration strategy based on the obtained matching degree to obtain a second transformer capacity configuration strategy of the target transformer area;

[0080] Specifically, step S3 includes:

[0081] 1) performing over-limit statistics on the real-time load data based on the transformer load rate threshold, and performing normalization processing on the statistical results to obtain a transformer capacity matching degree quantization value;

[0082] In actual operation of the target transformer area, the load fluctuation amplitude of adjacent time points is within 15%, indicating that the load change is relatively stable. Therefore, the transformer load rate threshold in this embodiment is set to 85%. Based on the real-time load data, the real-time load rate is calculated, and when the real-time load rate exceeds the transformer load rate threshold, it is recorded as an over-limit. After the over-limit times are counted, the transformer capacity matching degree quantization value is obtained after normalization processing.

[0083] 2) calculating the transformer capacity margin coefficient corresponding to the first transformer capacity configuration strategy based on support vector regression;

[0084] In the calculation of the transformer capacity margin coefficient, factors related to the transformer capacity margin coefficient, such as the load fluctuation amplitude, the transformer load rate and the transformer capacity matching degree quantization value, are first taken as input features, and the known transformer capacity margin coefficient is taken as an output label, and then a regression model is constructed in these data by using the support vector regression algorithm. The regression model can calculate the corresponding transformer capacity margin coefficient according to the input first transformer capacity configuration strategy.

[0085] 3) performing weighted summation on the transformer capacity matching degree quantization value and the transformer capacity margin coefficient to obtain the matching degree of the real-time load data and the first transformer capacity configuration strategy;

[0086] The analytic hierarchy process is used to compare the relative importance between the transformer capacity matching quantization value and the transformer capacity margin coefficient, and the weight coefficient is calculated by constructing a judgment matrix. Based on the weight coefficient, the transformer capacity matching quantization value and the transformer capacity margin coefficient are weighted and summed to obtain the matching degree of the real-time load data and the first transformer capacity configuration strategy.

[0087] 4) If the matching degree is not higher than the first preset threshold, the first transformer capacity configuration strategy is iteratively optimized to obtain the second transformer capacity configuration strategy of the target transformer area.

[0088] The first preset threshold of the embodiment is 0.85. When the matching degree is lower than 0.85, the first transformer capacity configuration strategy is iteratively optimized to obtain the second transformer capacity configuration strategy of the target transformer area.

[0089] Specifically, the gradient descent method is used to iteratively optimize the first transformer capacity configuration strategy until the first convergence condition is reached to obtain the second transformer capacity configuration strategy of the target transformer area. Further, the iterative optimization includes adjusting the transformer capacity level and the transformer capacity margin coefficient. The adjustment of the transformer capacity level includes increasing or decreasing the transformer capacity to better adapt to the load change; the adjustment of the transformer capacity margin coefficient is to change the proportion of the reserved capacity to improve the future load coping capacity. After adjustment, the matching degree of the real-time load data and the adjusted first transformer capacity configuration strategy is calculated again to obtain a new matching degree. If the new matching degree still does not meet the requirements, the adjustment and calculation are continued, and through continuous iterative optimization, the first convergence condition is reached, that is, the difference between the matching degrees of two consecutive iterations is less than the second preset threshold.

[0090] When the gradient descent method is used for iterative optimization in the embodiment, the learning rate is 0.01 and the second preset threshold is 0.01. When the difference between the matching degrees of two consecutive iterations is less than 0.01, it is determined that the convergence is reached, and the second transformer capacity configuration strategy is output. The capacity configuration strategy can meet the current electricity demand and also reserves a reasonable margin for future load growth.

[0091] S4, obtaining real-time operation data of each transformer in the target transformer area, and obtaining a transformer cooperative control strategy of the target transformer area based on the real-time operation data using a particle swarm optimization algorithm.

[0092] Specifically, step S4 includes:

[0093] 1) Based on the real-time operation data, a first transformer cooperative control strategy of the target transformer area is obtained;

[0094] The first transformer cooperative control strategy includes the real-time active power, real-time reactive power and real-time voltage of each transformer.

[0095] 2) The particle swarm optimization algorithm is used to iteratively optimize the cooperative control strategy of the first transformer until the second convergence condition is met, so as to obtain the cooperative control strategy of the second transformer in the target area.

[0096] In particle swarm optimization, particle positions correspond to transformer operating data, including active power, reactive power, and voltage. The fitness function uses load balance as the optimization objective; the difference between the maximum and minimum transformer load rates within the target transformer area represents the load imbalance of the target area.

[0097] Specifically, iterative optimization includes calculating the load imbalance degree of the target transformer area based on the load rate of each transformer. In this embodiment, when the load imbalance degree is less than 0.2, the target transformer area is determined to have reached a balanced state. The second convergence condition is that the difference in load imbalance degree over three consecutive iterations is less than a third preset threshold. In this embodiment, the third preset threshold is 0.05. When the difference in load imbalance degree over three consecutive iterations, i.e., the fitness increment, is less than 0.05, convergence is determined, and the second transformer collaborative control strategy is output.

[0098] 3) The second transformer collaborative control strategy is characterized as the target transformer collaborative control strategy.

[0099] The transformer coordinated control strategy includes the active power regulation target value, reactive power regulation target value, and voltage regulation target value for each transformer.

[0100] This invention discloses a method for transformer capacity configuration and collaborative control in a distribution network. By analyzing historical load data, it can predict load density change trends. Combining semantic analysis and multi-step modeling, a first transformer capacity configuration strategy is generated. By introducing the matching degree calculation between real-time load data and the first transformer capacity configuration strategy, the transformer capacity configuration can be optimized. For each transformer, a collaborative control strategy is formulated using a particle swarm optimization algorithm, realizing intelligent configuration and group collaborative control of transformer capacity, thereby improving the operating efficiency and reliability of the distribution network.

[0101] like Figure 2 The diagram shown is a structural schematic of a power distribution network transformer capacity configuration and coordinated control system according to an embodiment of the present invention. (Refer to...) Figure 2 An embodiment of the present invention provides a distribution network transformer capacity configuration and coordinated control system, comprising:

[0102] The decomposition and prediction module 1 is used to acquire historical load data of the target distribution area in the distribution network, and to decompose and predict the historical load data to determine the load density change trend of the target distribution area.

[0103] Specifically, the decomposition prediction module includes:

[0104] a multi-scale decomposition unit, configured to perform multi-scale decomposition on the historical load data by using a discrete wavelet transform to obtain a multi-time granularity load data sequence;

[0105] a load density prediction unit, configured to perform data extraction on the multi-time granularity load data sequence by using a first recurrent neural network, and perform load density prediction based on at least the obtained time feature vector to obtain load density prediction data of the target transformer area;

[0106] a load density change trend determination unit, configured to cluster the load density prediction data, and identify a load density growth trend of the target transformer area based on a clustering result.

[0107] a first capacity configuration strategy determination module 2, configured to perform semantic analysis on the load density change trend, and perform multi-step modeling on obtained load semantic features to obtain a first transformer capacity configuration strategy of the target transformer area, wherein the multi-step modeling includes intra-day load feature matrix modeling, transformer capacity prediction modeling, and transformer capacity configuration modeling;

[0108] Specifically, the first capacity configuration strategy determination module includes:

[0109] a semantic analysis unit, configured to perform semantic analysis on the load density change trend by using a large language model to obtain load semantic features;

[0110] an intra-day load feature matrix determination unit, configured to perform over-limit statistics on historical load data based on the load semantic features, and construct an intra-day load feature matrix of the target transformer area according to a statistical result, wherein the intra-day load feature matrix includes load features at a plurality of time points in a day, and the load features include peak load, duration, and occurrence period;

[0111] a transformer capacity level determination unit, configured to perform capacity mapping on the intra-day load feature matrix by using a second recurrent neural network, and cluster obtained transformer capacity prediction data to obtain a plurality of transformer capacity levels;

[0112] a priority ranking unit, configured to perform priority ranking on each transformer capacity level based on an analytic hierarchy process, and represent a transformer capacity level with the highest priority as the first transformer capacity configuration strategy of the target transformer area.

[0113] a second capacity configuration strategy determination module 3, configured to obtain real-time load data of the target transformer area, perform matching degree calculation on the real-time load data and the first transformer capacity configuration strategy, and optimize the first transformer capacity configuration strategy based on obtained matching degrees to obtain a second transformer capacity configuration strategy of the target transformer area;

[0114] Specifically, the second capacity configuration strategy determination module includes:

[0115] The capacity matching quantification value determination unit is configured to perform over-limit statistics on the real-time load data based on a transformer load rate threshold, and perform normalization processing on the statistical result to obtain a transformer capacity matching quantification value.

[0116] The capacity margin coefficient determination unit is configured to calculate a transformer capacity margin coefficient corresponding to the first transformer capacity configuration strategy based on support vector regression.

[0117] The matching degree calculation unit is configured to perform weighted summation on the transformer capacity matching quantification value and the transformer capacity margin coefficient to obtain a matching degree of the real-time load data and the first transformer capacity configuration strategy.

[0118] The iterative optimization unit is configured to determine whether the matching degree is higher than a first preset threshold, and if not, perform iterative optimization on the first transformer capacity configuration strategy to obtain a second transformer capacity configuration strategy of the target transformer area.

[0119] The coordinated control strategy determination module 4 is configured to obtain real-time operation data of each transformer in the target transformer area, and obtain a transformer coordinated control strategy of the target transformer area based on the real-time operation data using a particle swarm optimization algorithm, wherein the transformer coordinated control strategy includes an active power regulation target value, a reactive power regulation target value and a voltage regulation target value of each transformer.

[0120] It should be noted that the above-mentioned various modules in the power distribution network transformer capacity configuration and coordinated control system can be realized by software, hardware and combinations thereof. The above-mentioned various modules can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory in the computer device in software form, so as to be called and executed by the processor to perform the operations corresponding to the above-mentioned various modules. For specific limitations of the power distribution network transformer capacity configuration and coordinated control system, refer to the limitations of the power distribution network transformer capacity configuration and coordinated control method, both of which have the same functions and effects, and will not be described here.

[0121] In summary, the power distribution network transformer capacity configuration and coordinated control method and system can predict the load density trend by analyzing historical load data, generate a first transformer capacity configuration strategy by combining semantic analysis and multi-step modeling, introduce the matching degree calculation of real-time load data and the first transformer capacity configuration strategy, optimize the transformer capacity configuration, use the particle swarm optimization algorithm to formulate a coordinated control strategy for each transformer, realize intelligent configuration and group coordinated control of transformer capacity, and improve the operation efficiency and reliability of the power distribution network.

[0122] The various embodiments in this specification are described in a progressive manner. For directly identical or similar parts of the embodiments, refer to each other. Each embodiment focuses on its differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments. It should be noted that the technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0123] 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 substitutions can be made without departing from the technical principles of the present invention, and these improvements and substitutions should also be considered within the scope of protection of the present invention.

Claims

1. A method for transformer capacity configuration and coordinated control in a distribution network, characterized in that, include: Historical load data of the target distribution area in the distribution network is obtained, and the historical load data is decomposed and predicted to determine the load density change trend of the target distribution area. Semantic analysis is performed on the load density change trend, and multi-step modeling is performed on the obtained load semantic features to obtain the first transformer capacity configuration strategy for the target distribution area. The multi-step modeling includes intraday load feature matrix modeling, transformer capacity prediction modeling, and transformer capacity configuration modeling. The real-time load data of the target transformer area is obtained, the matching degree between the real-time load data and the first transformer capacity configuration strategy is calculated, and the first transformer capacity configuration strategy is optimized based on the obtained matching degree to obtain the second transformer capacity configuration strategy of the target transformer area. The real-time operating data of each transformer in the target distribution area is obtained, and a particle swarm optimization algorithm is used to obtain a transformer collaborative control strategy for the target distribution area based on the real-time operating data. The transformer collaborative control strategy includes the active power regulation target value, reactive power regulation target value and voltage regulation target value for each transformer. The process of performing semantic analysis on the load density change trend and multi-step modeling on the obtained load semantic features to obtain the first transformer capacity configuration strategy for the target distribution area includes: A large language model was used to perform semantic analysis on the load density change trend to obtain load semantic features; Based on the load semantic features, the historical load data is statistically analyzed for exceeding limits, and a daily load feature matrix of the target transformer area is constructed based on the statistical results. The daily load feature matrix includes load features at several time points throughout the day, and the load features include peak load, duration, and occurrence time. A second recurrent neural network is used to perform capacity mapping on the intraday load feature matrix, and the obtained transformer capacity prediction data is clustered to obtain several transformer capacity levels. The priority of each transformer capacity level is sorted based on the analytic hierarchy process, and the transformer capacity level with the highest priority is represented as the first transformer capacity configuration strategy for the target distribution area.

2. The method for distribution network transformer capacity configuration and coordinated control according to claim 1, characterized in that, The step of decomposing and predicting the historical load data to determine the load density change trend of the target distribution area includes: The historical load data is decomposed into multiple scales using discrete wavelet transform to obtain a multi-time granularity load data sequence. The first recurrent neural network is used to extract data from the multi-time granularity load data sequence, and load density prediction is performed based at least on the obtained time feature vector to obtain the load density prediction data of the target area. The load density prediction data is clustered, and the load density growth trend of the target area is identified based on the clustering results.

3. The method for distribution network transformer capacity configuration and coordinated control according to claim 1, characterized in that, The step of calculating the matching degree between the real-time load data and the first transformer capacity configuration strategy, and optimizing the first transformer capacity configuration strategy based on the obtained matching degree to obtain the second transformer capacity configuration strategy for the target distribution area includes: The real-time load data is statistically analyzed based on the transformer load rate threshold, and the statistical results are normalized to obtain a quantified value for transformer capacity matching. The transformer capacity margin coefficient corresponding to the first transformer capacity configuration strategy is calculated based on support vector regression. The matching degree between the real-time load data and the first transformer capacity configuration strategy is obtained by weighted summation of the quantified value of the transformer capacity matching metric and the transformer capacity margin coefficient. Determine whether the matching degree is higher than a first preset threshold. If not, iteratively optimize the first transformer capacity configuration strategy to obtain the second transformer capacity configuration strategy for the target area.

4. The method for distribution network transformer capacity configuration and coordinated control according to claim 3, characterized in that, The iterative optimization of the first transformer capacity configuration strategy to obtain the second transformer capacity configuration strategy for the target distribution area includes: The gradient descent method is used to iteratively optimize the first transformer capacity configuration strategy until the first convergence condition is met, thereby obtaining the second transformer capacity configuration strategy for the target transformer area. The iterative optimization includes adjusting the transformer capacity level and the transformer capacity margin coefficient. The first convergence condition is that the difference in matching degree between two consecutive iterations is less than a second preset threshold.

5. The method for distribution network transformer capacity configuration and coordinated control according to claim 1, characterized in that, The transformer collaborative control strategy for the target distribution area, obtained using a particle swarm optimization algorithm based on the real-time operating data, includes: Based on the real-time operating data, a first transformer collaborative control strategy for the target distribution area is obtained, wherein the first transformer collaborative control strategy includes the real-time active power, real-time reactive power and real-time voltage of each transformer. The first transformer collaborative control strategy is iteratively optimized using a particle swarm optimization algorithm until a second convergence condition is met, thereby obtaining a second transformer collaborative control strategy for the target transformer area. The iterative optimization includes calculating the load imbalance degree of the target transformer area based on the load rate of each transformer. The second convergence condition is that the difference in load imbalance degree for three consecutive iterations is less than a third preset threshold. The second transformer collaborative control strategy is characterized as the transformer collaborative control strategy of the target transformer area.

6. A distribution network transformer capacity configuration and coordinated control system, characterized in that, include: The decomposition and prediction module is used to acquire historical load data of the target distribution area in the distribution network, and to decompose and predict the historical load data to determine the load density change trend of the target distribution area. The first capacity configuration strategy determination module is used to perform semantic analysis on the load density change trend and perform multi-step modeling on the obtained load semantic features to obtain the first transformer capacity configuration strategy for the target area. The multi-step modeling includes intraday load feature matrix modeling, transformer capacity prediction modeling, and transformer capacity configuration modeling. The second capacity configuration strategy determination module is used to acquire real-time load data of the target transformer area, calculate the matching degree between the real-time load data and the first transformer capacity configuration strategy, and optimize the first transformer capacity configuration strategy based on the obtained matching degree to obtain the second transformer capacity configuration strategy of the target transformer area. The collaborative control strategy determination module is used to acquire real-time operating data of each transformer in the target distribution area, and to obtain the collaborative control strategy of the transformers in the target distribution area based on the real-time operating data using a particle swarm optimization algorithm. The collaborative control strategy of the transformers includes the active power regulation target value, reactive power regulation target value and voltage regulation target value of each transformer. The first capacity configuration strategy determination module includes: The semantic analysis unit is used to perform semantic analysis on the load density change trend using a large language model to obtain load semantic features; The intraday load feature matrix determination unit is used to perform over-limit statistics on the historical load data based on the load semantic features, and construct the intraday load feature matrix of the target transformer area based on the statistical results. The intraday load feature matrix includes load features at several time points in a day, and the load features include peak load, duration and occurrence time. The transformer capacity level determination unit is used to perform capacity mapping on the intraday load feature matrix using a second recurrent neural network, and to cluster the obtained transformer capacity prediction data to obtain several transformer capacity levels. The priority sorting unit is used to prioritize each transformer capacity level based on the analytic hierarchy process, and to characterize the transformer capacity level with the highest priority as the first transformer capacity configuration strategy for the target distribution area.

7. The distribution network transformer capacity configuration and coordinated control system according to claim 6, characterized in that, The decomposition prediction module includes: A multi-scale decomposition unit is used to perform multi-scale decomposition on the historical load data using discrete wavelet transform to obtain a multi-time granularity load data sequence. The load density prediction unit is used to extract data from the multi-time granularity load data sequence using a first recurrent neural network, and to predict the load density based at least on the obtained time feature vector, so as to obtain the load density prediction data of the target area. The load density change trend determination unit is used to cluster the load density prediction data and identify the load density growth trend of the target area based on the clustering results.

8. The distribution network transformer capacity configuration and coordinated control system according to claim 6, characterized in that, The second capacity configuration strategy determination module includes: The capacity matching quantification value determination unit is used to perform over-limit statistics on the real-time load data based on the transformer load rate threshold, and to normalize the statistical results to obtain the transformer capacity matching quantification value. The capacity margin coefficient determination unit is used to calculate the transformer capacity margin coefficient corresponding to the first transformer capacity configuration strategy based on support vector regression. The matching degree calculation unit is used to perform a weighted summation of the transformer capacity matching quantified value and the transformer capacity margin coefficient to obtain the matching degree between the real-time load data and the first transformer capacity configuration strategy. An iterative optimization unit is used to determine whether the matching degree is higher than a first preset threshold. If not, iterative optimization is performed on the first transformer capacity configuration strategy to obtain a second transformer capacity configuration strategy for the target transformer area.

Images