A power distribution network planning method adaptive to distributed new energy access

Abstract

The application discloses a power distribution network planning method suitable for distributed new energy access, and relates to the technical field of power distribution networks.The method comprises the following steps: calculating the data feature weight of a power distribution network sub-area by an entropy weight method; applying a KMEANS algorithm to cluster data to generate a power distribution network sub-area cluster, and combining a cluster center and a feature weight to calculate a weighted cluster center; using a sorting method close to an ideal solution to score the weighted cluster center, and dividing the function label of the power distribution network sub-area; calculating the channel transmission capacity index and the stability index of the power distribution network sub-area, combining the obtained channel comprehensive index, and classifying the channel as a main road or a redundant channel; constructing a topological atlas based on the function label and the channel category, counting the number of main road connections and the number of redundant channel connections of the sub-area, calculating the energy storage index of the power distribution network sub-area by combining the function label, and deploying a centralized or edge energy storage unit; integrating adjustable load data and the energy storage unit, constructing an energy storage load collaborative model, and realizing dynamic power consumption planning of the power distribution network with new energy access.

Description

Technical Field

[0001] This invention relates to the field of power distribution network technology, specifically to a power distribution network planning method adapted to the integration of distributed renewable energy sources. Background Technology

[0002] With the rapid development of renewable energy, especially the large-scale integration of distributed energy sources such as photovoltaics and wind power, the operation of power systems faces new challenges. Traditional distribution networks are typically planned based on static load and stability assumptions. However, with the integration of new energy sources, the volatility and uncertainty of power loads increase significantly, adding complexity to grid dispatching and management. To ensure the safe and reliable operation of the grid, the planning and optimization of distribution networks need to be more refined, especially in terms of how to effectively integrate and dispatch distributed energy sources, energy storage devices, and adjustable loads, requiring more advanced technologies and methods.

[0003] Currently, many existing technical solutions rely on simplified load forecasting and static analysis methods, which are based solely on the statistical average of historical data. This results in a lack of accurate quantitative description and real-time response mechanism for the dynamic changes in power load, and fails to effectively capture the dynamic fluctuation characteristics of new energy sources and adjust the power grid accordingly. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a distribution network planning method adapted to distributed renewable energy access, thereby solving the problems mentioned in the background section.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a distribution network planning method adapted to distributed renewable energy access, comprising the following steps:

[0006] Step S1: Collect and preprocess the operation data of the distribution network sub-area to obtain the operation dataset of the distribution network sub-area. Calculate the entropy value of the operation dataset of the distribution network sub-area using the entropy weight method to obtain the feature weight of each feature column of the operation dataset of the distribution network sub-area.

[0007] Step S2: Cluster the power distribution sub-area operation dataset using the KMEANS algorithm to obtain power distribution sub-area clusters; calculate the weighted cluster center by combining the cluster center of the power distribution sub-area clusters with the feature weight of each feature column of the power distribution sub-area operation dataset;

[0008] Step S3: Calculate the scores of the weighted cluster centers using a sorting method that approximates the ideal solution to obtain the maximum approximation score of the weighted cluster centers. Then, use the maximum approximation score to divide the distribution network sub-areas into functional labels to obtain the functional labels of the distribution network sub-areas.

[0009] Step S4: Collect distribution network channel data, including voltage level, feeder capacity, voltage drop probability, and short-time frequency response capability; calculate the channel transmission capacity index by combining voltage level and feeder capacity; calculate the channel stability index by combining voltage drop probability and short-time frequency response capability; calculate the channel comprehensive index by combining the channel transmission capacity index and channel stability index; classify the channels in the distribution network data using the channel comprehensive index to obtain channel categories, namely: trunk channels and redundant channels; construct a distribution network channel topology map using the functional labels of the distribution network sub-areas and the channel categories.

[0010] Step S5: Based on the distribution network channel topology map, obtain the number of main road connections and redundant channel connections of the distribution network sub-area; by combining the functional tags of the distribution network sub-area, the number of main road connections and redundant channel connections of the distribution network sub-area, calculate the energy storage index of the distribution network sub-area; deploy energy storage units for the distribution network sub-area based on the energy storage index of the distribution network sub-area.

[0011] Step S6: Collect adjustable load data between sub-sections of the distribution network, construct an energy storage load coordination model based on the adjustable load data between the sub-sections of the distribution network and the energy storage unit, and realize the power consumption planning of the distribution network for new energy access through the energy storage load coordination model.

[0012] Preferably, the step of calculating the entropy value of the distribution network operation dataset using the entropy weight method to obtain the feature weight of each feature column of the distribution network operation dataset includes the following specific steps:

[0013] The entropy value of the characteristic column in the sub-area operation data of the distribution network is calculated using the entropy weight method:

[0014]

[0015] in, is the entropy value of the j-th feature in the distribution network sub-area operation data, and n is the number of samples in the distribution network sub-area operation data. This refers to the proportion of data from the i-th distribution network sub-area in the j-th feature column, where j represents the index of the j-th feature and i represents the data from the i-th distribution network sub-area.

[0016] Calculate the weight of each feature in the sub-area operation data of the distribution network:

[0017]

[0018] in, is the weight of the j-th feature of the distribution network sub-area operation data, where J is the total number of features of the distribution network sub-area operation data. It is the entropy value of the j-th feature of the distribution network sub-area operation data, where j represents the index of the j-th feature.

[0019] Preferably, the step of calculating the weighted cluster center by combining the cluster centers of the distribution network sub-area clusters and the feature weights of each feature column of the distribution network sub-area operation dataset includes the following steps:

[0020] By combining the cluster centers of the distribution network sub-area clusters with the aforementioned feature weights, the weighted cluster centers are calculated:

[0021]

[0022] in, This represents the value of the k-th cluster center after weighting the j-th feature. This represents the feature weight of the j-th feature. Indicates the first sub-region cluster of the distribution network The value of the cluster center on the j-th feature, where j represents the index of the j-th feature, and K represents the total number of cluster centers.

[0023] Preferably, the step of calculating the scores of the weighted cluster centers using a sorting method that approximates the ideal solution to obtain the maximum approximation score of the weighted cluster centers includes the following steps:

[0024] For a distribution network sub-area with function label L, its positive ideal solution is: The negative ideal solution is ;

[0025] For each distribution network sub-area cluster, calculate the distance between its weighted cluster center and the positive ideal solution for different functional labels of the distribution network sub-area:

[0026]

[0027] in, The distance between the k-th distribution network sub-area cluster and the positive ideal solution of the L-th functional label, where j represents the index of the j-th feature, and J is the total number of features in the distribution network sub-area operational data. This represents the value of the k-th cluster center after weighting the j-th feature. This represents the value of the positive ideal solution of the Lth functional label in the jth feature;

[0028] For each distribution network sub-region cluster, calculate the distance between its weighted cluster center and the negative ideal solution of different functional labels of the sub-region:

[0029]

[0030] in, The distance between the k-th distribution network sub-area cluster and the negative ideal solution of the L-th functional label, where j represents the index of the j-th feature, and J is the total number of features in the distribution network sub-area operational data. This represents the value of the k-th cluster center after weighting the j-th feature. This represents the value of the negative ideal solution of the Lth functional label in the jth feature;

[0031] For each distribution network sub-area cluster, the maximum approximation score of the cluster is calculated by using the distances between its weighted cluster center and the positive ideal solutions of different functional labels of the sub-area.

[0032]

[0033] in, Indicates the first The maximum approximation score of each cluster reflects how close it is to the ideal solution. The distance between the k-th distribution network sub-area cluster and the positive ideal solution of the L-th functional label. The distance between the k-th distribution network sub-area cluster and the negative ideal solution of the L-th functional label is given by max(), which is a maximum value function representing the maximum approximation score between the k-th distribution network sub-area cluster and the L functional labels. This represents the similarity score between the k-th distribution network sub-area cluster and the L-th functional label.

[0034] Preferably, the step of dividing the distribution network sub-area into functional labels based on the maximum approximation score to obtain the functional labels of the distribution network sub-area includes the following specific steps:

[0035] For the k-th distribution network sub-area cluster, by calculating the approximation score between its weighted cluster center and different functional labels, the maximum approximation score between the k-th distribution network sub-area cluster and the L-th functional label is taken as the maximum approximation score of the k-th distribution network sub-area cluster. Therefore, the functional label of the k-th distribution network sub-area cluster is... .

[0036] Preferably, the calculation of the channel transmission capacity index by combining voltage level and feeder capacity includes the following specific steps:

[0037] The channel transmission capacity index is calculated by combining the voltage level and feeder capacity:

[0038]

[0039] in, V represents the channel transmission capacity index, where V is the voltage level. Where F is the highest voltage of the channel and F is the feeder capacity. This represents the maximum feeder capacity of the channel.

[0040] Preferably, the calculation of the channel stability index by combining voltage drop probability and short-time frequency response capability includes the following specific steps:

[0041] The channel stability index is calculated by combining the voltage drop probability and short-time frequency response capability:

[0042]

[0043] in, Here, D is the channel stability index, D is the voltage sag probability, and Q is the short-time frequency response capability. This represents the maximum probability of voltage drop. This represents the maximum short-time frequency response capability.

[0044] Preferably, the calculation of the comprehensive channel index by combining the channel transmission capacity index and the channel stability index includes the following specific steps:

[0045] The overall channel index is calculated by combining the channel transmission capacity index and the channel stability index:

[0046]

[0047] Wherein, ZN is the channel composite index. This is the channel transmission capacity index. This is the channel stability index.

[0048] Preferably, the step of calculating the energy storage index of a distribution network sub-area by combining the functional tags of the sub-area, the number of main road connections and the number of redundant channel connections of the sub-area includes the following specific steps:

[0049] Based on the aforementioned distribution network channel topology map, and by combining the functional tags of the distribution network sub-regions, the number of main channel connections and the number of redundant channel connections in the distribution network sub-regions, the energy storage index of the distribution network sub-regions is calculated:

[0050]

[0051] in, This represents the energy storage index of the i-th distribution network sub-region. This represents the number of main road connections in the i-th distribution network sub-area. The functional label represents the cluster to which the i-th distribution network sub-area belongs, and This represents the number of redundant channel connections in the i-th distribution network sub-area. and is a weighting coefficient used to smooth the importance of the main channel and the redundant channel, and a and b are sublinear exponents used to adjust the impact of the main channel and the redundant channel on the energy storage configuration.

[0052] Preferably, the construction of the energy storage load coordination model includes the following specific steps:

[0053] Construct an energy storage load coordination model, the objective function of which is:

[0054]

[0055] in, Let be the objective function. To optimize the total number of time periods within the time window, n is the number of samples of operational data for the distribution network sub-area. The total number of adjustable loads, For the i-th energy storage device during the time period Operating costs are related to time-of-use electricity pricing. For adjustable load m during time period The power adjustment amount is positive for increasing the load and negative for reducing the load. The adjustment compensation unit price for adjustable load m;

[0056] Calculate the operating cost of energy storage equipment based on the different types of energy storage units:

[0057]

[0058] in, For the i-th energy storage device during the time period Operating costs This represents the charging power of the i-th energy storage device during time period t. This represents the discharge power of the i-th energy storage device during time period t. , This represents the charging cost coefficient and discharging cost coefficient of a centralized energy storage system during time period t. (t), This represents the charging cost coefficient and discharging cost coefficient of the edge energy storage unit in time period t;

[0059] The constraints of the energy storage load coordination model are: , , ,in, and This indicates the minimum and maximum allowable state of charge (SOC) of the energy storage device. The state of charge (SOC) of the i-th energy storage device during time period t represents the proportion of the energy currently stored by the device to its rated capacity, typically ranging from 0 to 1. Let m be the minimum power adjustment allowed for the adjustable load of type m during time period t. This represents the maximum allowable power adjustment for the m-th type of adjustable load during time period t. This represents the actual power adjustment of the m-th type of adjustable load during time period t.

[0060] Power constraints: ,in, This represents the total output power of new energy sources within time period t, which is the sum of the real-time power generation of distributed new energy sources. Let be the discharge power of the i-th energy storage device in time period t, that is, the power released by the energy storage to the grid. The rigid load power during time period t, This is the reference power for the m-th type of adjustable load. Let be the power adjustment amount of the m-th type of adjustable load during time period t. Let represent the charging power of the i-th energy storage device during time period t, where i is the i-th energy storage device, n is the number of samples of the distribution network sub-area operation data, m is the m-th type of adjustable load, and M is the total number of adjustable loads.

[0061] Beneficial effects

[0062] This invention provides a distribution network planning method adapted to distributed renewable energy access. It has the following advantages:

[0063] (1) By combining the cluster centers of the distribution network sub-areas with the feature weights of each feature column in the operational dataset, the functional division of distribution network sub-areas can be made more accurate. This method calculates the feature weights based on the degree of variation of the data itself, making planning decisions more objective and accurate. For example, the entropy weight method can automatically identify features closely related to new energy fluctuations, helping planners focus on key variables, thereby improving the rationality of energy storage configuration and scheduling strategies, and ensuring the efficiency and reliability of the distribution network in actual operation.

[0064] (2) Combining the channel transmission capacity index and the channel stability index, the calculated comprehensive channel index helps to comprehensively evaluate the transmission capacity and stability of distribution network channels. This comprehensive index not only reflects the basic transmission capacity of the channel, but also considers the stability factors of voltage drop and frequency response capability, effectively improving the accuracy of channel classification. In distribution network planning, the use of the comprehensive index can help identify trunk lines and redundant channels, thereby optimizing the topology of the distribution network and ensuring the load balance and power transmission stability of channels under different operating conditions.

[0065] (3) By combining the functional labels, mainline connection quantity, and redundant channel connection quantity of the distribution network sub-area, the energy storage index of the distribution network sub-area is calculated, which can provide a scientific basis for the deployment of energy storage units. This method optimizes the configuration of the energy storage system by considering the connection characteristics and functional requirements of the sub-area, enabling energy storage devices to better match the operational needs of the distribution network. The energy storage index not only helps to identify the deployment methods of centralized energy storage systems and edge energy storage units, but also improves the flexibility of energy dispatch, enhances the risk resistance and emergency dispatch capabilities of the distribution network, and helps to achieve stable operation and efficient utilization after the access of new energy sources. Attached Figure Description

[0066] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0067] Figure 1 This is a flowchart of the steps of a distribution network planning method adapted to distributed new energy access proposed in this invention;

[0068] Figure 2 This is a step-by-step diagram of a distribution network planning method adapted to distributed renewable energy access proposed in this invention; Detailed Implementation

[0069] 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.

[0070] Please see Figure 1-2 The present invention provides a technical solution: a distribution network planning method adapted to distributed new energy access.

[0071] Step S1: Collect and preprocess the distribution network operation data to obtain the distribution network operation dataset. Calculate the entropy value of the distribution network operation dataset using the entropy weight method to obtain the feature weight of each feature column in the distribution network operation dataset.

[0072] Data collection and preprocessing of distribution network sub-area operation data are performed by acquiring raw data from power grid operation monitoring systems such as SCADA, smart meters, renewable energy power plants such as photovoltaic inverters and wind power controllers, and meteorological monitoring platforms. The data includes power values ​​recorded in 15-minute or hourly granularity from historical load curves, real-time and predicted renewable energy output (such as photovoltaic / wind power), grid parameters (feeder capacity, voltage level, line impedance, etc.), voltage and current monitoring records, fault logs, and weather information (solar intensity, wind speed, and temperature). After acquisition, the data is cleaned to remove missing values ​​(e.g., through linear interpolation) or invalid periods and outliers. Subsequently, the multi-source data is time-aligned and formatted to ensure consistent timestamps across data points. Next, standardization is performed to normalize features of different dimensions, such as megawatt-level load and percentage voltage drop, to the same range (e.g., 0-1), avoiding interference from dimensional differences in subsequent analysis. Finally, key features are extracted based on the research objectives: load peak-to-valley ratio, photovoltaic output volatility, and voltage stability coefficient. Redundant information is removed, resulting in the final distribution network sub-area operation dataset.

[0073] It should be noted that a distribution network sub-area refers to the smallest analytical unit in a distribution network with an independent data acquisition unit. For example: feeder power supply range: the area covered by a single feeder or a group of electrically connected feeders; electrical node group: a local power grid separated by switches or transformers; data acquisition node: a monitoring area covered by a smart meter or PMU synchronous phasor measurement device.

[0074] It should be noted that the sample data in the distribution network sub-area operation dataset consists of operation data from different energy storage stations, and does not include duplicate operation data from energy storage stations.

[0075] The entropy value of the distribution network sub-area operation dataset is calculated using the entropy weight method to obtain the feature weights of the distribution network sub-area operation dataset.

[0076] Entropy weighting is a multi-index weighting method based on information entropy theory, used to objectively evaluate the importance of each feature in the decision-making process. This method quantifies the uncertainty of each feature by calculating its information entropy; the smaller the entropy value, the greater the information content of the feature and the greater its impact on the decision result, thus assigning it a higher weight. Entropy weighting avoids subjective bias and is widely used in multi-attribute decision-making, feature selection, and model optimization.

[0077] The entropy value of the characteristic column in the sub-area operation data of the distribution network is calculated using the entropy weight method:

[0078]

[0079] in, is the entropy value of the j-th feature in the distribution network sub-area operation data, and n is the number of samples in the distribution network sub-area operation data. This refers to the proportion of data from the i-th distribution network sub-area in the j-th feature column, where j represents the index of the j-th feature and i represents the data from the i-th distribution network sub-area.

[0080] Calculate the weight of each feature in the sub-area operation data of the distribution network:

[0081]

[0082] in, is the weight of the j-th feature of the distribution network sub-area operation data, where J is the total number of features of the distribution network sub-area operation data. It is the entropy value of the j-th feature of the distribution network sub-area operation data, where j represents the index of the j-th feature.

[0083] It is important to note that in distribution network planning, calculating entropy values ​​from the operational dataset of distribution network sub-areas using the entropy weight method to obtain feature weights has several significant implications and benefits. Distribution network sub-area operational data contains numerous features, such as load characteristics, renewable energy access status, and line parameters, each with varying degrees of impact on planning. The entropy weight method can objectively determine the weights of these features, avoiding subjective arbitrariness. Traditional methods may assign weights to certain features based on experience, making it difficult to accurately reflect the actual situation. The entropy weight method, however, calculates weights based on the degree of variation in the data itself. Features with high variation carry more effective information and have a more critical impact on planning decisions; therefore, the entropy weight method assigns higher weights to such features. For example, in areas with high renewable energy penetration, the variation in renewable energy output fluctuations is significant. The entropy weight method emphasizes the weight of this feature, enabling planners to consider the impact of renewable energy fluctuations more carefully when formulating energy storage configuration and dispatch strategies, thereby improving the alignment between planning and actual operation.

[0084] Step S2: Cluster the power distribution sub-area operation dataset using the KMEANS algorithm to obtain power distribution sub-area clusters; calculate the weighted cluster center by combining the cluster center of the power distribution sub-area clusters with the feature weight of each feature column of the power distribution sub-area operation dataset.

[0085] The specific steps for clustering the sub-regional operation dataset of the distribution network using the KMEANS algorithm are as follows:

[0086] The number of clusters is determined based on the feature dimensions of the preprocessed dataset (such as load peak-to-valley ratio, photovoltaic output fluctuation rate, voltage drop probability, etc.). The initial number of clusters can be selected using the elbow rule: the elbow rule is to plot the sum of squared distances of total samples corresponding to different numbers of clusters and select the value at the inflection point as the optimal number of clusters. After determining the number of clusters, the KMEANS algorithm model is trained on the standardized data. The cluster centers and sample assignments are updated iteratively until the change in the center point is less than a set threshold (e.g., 0.001) or the maximum number of iterations (e.g., 300 times) is reached. In each iteration, the Euclidean distance from each sample to each cluster center is calculated, and the sample is assigned to the cluster corresponding to the nearest center. The mean of the samples within the cluster is then recalculated as the new center.

[0087] After clustering is completed, the results need to be analyzed for validity. Visualization tools (such as PCA dimensionality reduction followed by a 2D scatter plot or parallel coordinate plot) should be used to observe whether the cluster distribution is reasonable and to check whether the differences in characteristics between clusters are significant. For example, the standard deviation or coefficient of variation of features within each cluster should be calculated to ensure that similar regions have similar characteristics. If excessive differences in sample characteristics are found within a cluster (e.g., voltage drop probability ranging from 5% to 25%), the number of clusters should be adjusted or the feature weights optimized.

[0088] The KMEANS algorithm was used to cluster the operational dataset of the distribution network sub-areas, resulting in the following cluster number for each sub-area: The cluster center of each cluster is:

[0089]

[0090] in, Let represent the k-th cluster center, and j represent the index of the j-th feature. Indicates the first sub-region cluster of the distribution network The value of the cluster center on the j-th feature, where K represents the total number of cluster centers.

[0091] By combining the cluster centers of the distribution network sub-area clusters with the aforementioned feature weights, the weighted cluster centers are calculated:

[0092]

[0093] in, This represents the value of the k-th cluster center after weighting the j-th feature. This represents the feature weight of the j-th feature. Indicates the first sub-region cluster of the distribution network The value of the cluster center on the j-th feature, where j represents the index of the j-th feature, and K represents the total number of cluster centers.

[0094] Step S3: Calculate the scores of the weighted cluster centers by using a sorting method that approximates the ideal solution to obtain the maximum approximation score of the weighted cluster centers. Then, use the maximum approximation score to divide the distribution network sub-regions into functional labels to obtain the functional labels of the distribution network sub-regions.

[0095] In distribution network planning, functional labels for distribution network sub-areas will be used. The load is classified into three types: significant peak load during the day, photovoltaic output mutation sensitive type, and industrial continuous load type. The positive and negative ideal solutions for the significant peak load during the day, photovoltaic output mutation sensitive type, and industrial continuous load type are determined by a sorting method that approximates the ideal solution.

[0096] It should be noted that the "Photovoltaic Output Fluctuation Sensitivity" label is primarily defined by output fluctuation sensitivity and the potential for new energy development planning. Output fluctuation sensitivity measures the impact of sudden changes in new energy (such as photovoltaic) output on the power grid. If a region's photovoltaic output is significantly affected by weather changes (e.g., cloud cover causing a sudden drop in power), and the region's power grid structure is sensitive to such fluctuations (e.g., limited feeder capacity, insufficient voltage regulation), then its output fluctuation sensitivity is high. Simultaneously, the potential for new energy development planning reflects the scale of future exploitable photovoltaic resources in the region. If a region has abundant solar resources and plans for a large amount of photovoltaic installations, even if current output fluctuations are small, the potential fluctuation risk cannot be ignored. Therefore, the combination of these two labels can identify areas that require energy storage or dynamic regulation devices to mitigate output fluctuations and ensure grid stability, such as rural areas or emerging energy development zones. The ideal solution for the "Photovoltaic Output Fluctuation Sensitivity" label is: =[Neutral value of load curve shape, maximum value of power output fluctuation sensitivity, neutral value of historical load peak-to-valley ratio, neutral value of load power factor, maximum value of renewable energy potential], negative ideal solution =[Neutral value of load curve shape, minimum value of power output fluctuation sensitivity, neutral value of historical load peak-to-valley ratio, neutral value of load power factor, minimum value of new energy potential]. Here, the neutral value represents the average value of this characteristic, and the maximum value represents the maximum value of this characteristic. For example, the neutral value of load curve shape represents the average value of the load curve shapes of k cluster centers.

[0097] It should be noted that the "Significant Daytime Load Peak" label is determined by the load curve shape and the historical peak-to-valley ratio. The load curve shape reflects the temporal distribution characteristics of electricity consumption. If the load curve of a certain area shows a significant peak during the day (e.g., from morning to evening) and a sharp drop in electricity consumption at night, it indicates that it has daytime peak characteristics. The historical peak-to-valley ratio further quantifies the intensity of the peak-to-valley difference. A high ratio means that the electricity consumption in the area fluctuates drastically, with significant peak-to-valley differences. For example, commercial or residential areas experience concentrated electricity consumption during the day and a sharp drop in electricity demand at night; their load curve shape and peak-to-valley ratio both conform to the characteristics of this label. Through the combination of these two labels, the system can identify areas that need to focus on addressing daytime power supply pressure and require energy storage to smooth out peaks and valleys. In the positive ideal solution of the "Significant Daytime Load Peak" label, the load curve shape and the historical peak-to-valley ratio take their maximum values, while other characteristics take neutral values. In the negative ideal solution of the "Significant Daytime Load Peak" label, the load curve shape and the historical peak-to-valley ratio take their minimum values, while other characteristics take neutral values.

[0098] It's important to note that the key to the "industrial-grade continuous load" label lies in the synergistic effect of the load power factor and the load curve shape. The load power factor reflects the energy utilization efficiency of electrical equipment. Industrial loads (such as motors and frequency converters) typically have a high power factor (close to 1), indicating high energy conversion efficiency and low reactive power demand from the grid. Furthermore, the load curve shape in industrial areas is usually relatively stable, without significant peak-to-valley fluctuations, as their production processes require continuous operation (e.g., 24-hour production lines). Therefore, the combination of a high power factor and low load fluctuations is a typical characteristic of the industrial-grade continuous load type. These areas have high requirements for grid stability and need to be equipped with energy storage to cope with possible short-term disturbances, but frequent adjustments are not required, such as in heavy industrial bases or chemical industrial parks. In the positive ideal solution of the industrial-grade continuous load type, the load power factor and load curve shape take their maximum values, while other characteristics take neutral values. Conversely, in the positive ideal solution of the industrial-grade continuous load type, the load power factor and load curve shape take their minimum values, while other characteristics take neutral values.

[0099] For a distribution network sub-area with function label L, its positive ideal solution is: The negative ideal solution is .

[0100] For each distribution network sub-area cluster, calculate the distance between its weighted cluster center and the positive ideal solution for different functional labels of the distribution network sub-area:

[0101]

[0102] in, The distance between the k-th distribution network sub-area cluster and the positive ideal solution of the L-th functional label, where j represents the index of the j-th feature, and J is the total number of features in the distribution network sub-area operational data. This represents the value of the k-th cluster center after weighting the j-th feature. Let represent the value of the positive ideal solution of the Lth functional label in the jth feature.

[0103] For each distribution network sub-region cluster, calculate the distance between its weighted cluster center and the negative ideal solution of different functional labels of the sub-region:

[0104]

[0105] in, The distance between the k-th distribution network sub-area cluster and the negative ideal solution of the L-th functional label, where j represents the index of the j-th feature, and J is the total number of features in the distribution network sub-area operational data. This represents the value of the k-th cluster center after weighting the j-th feature. Let represent the value of the negative ideal solution of the Lth functional label in the jth feature.

[0106] For each distribution network sub-area cluster, the maximum approximation score of the cluster is calculated by using the distances between its weighted cluster center and the positive and negative ideal solutions of different functional labels of the sub-area.

[0107]

[0108] in, Indicates the first The maximum approximation score of each cluster reflects how close it is to the ideal solution. The distance between the k-th distribution network sub-area cluster and the positive ideal solution of the L-th functional label. The distance between the k-th distribution network sub-area cluster and the negative ideal solution of the L-th functional label is given by max(), which is a maximum value function representing the maximum approximation score between the k-th distribution network sub-area cluster and the L functional labels. This represents the similarity score between the k-th distribution network sub-area cluster and the L-th functional label.

[0109] For the k-th distribution network sub-area cluster, by calculating the approximation score between its weighted cluster center and different functional labels, the maximum approximation score between the k-th distribution network sub-area cluster and the L-th functional label is taken as the maximum approximation score of the k-th distribution network sub-area cluster. Therefore, the functional label of the k-th distribution network sub-area cluster is... .

[0110] Step S4: Collect distribution network channel data, including voltage level, feeder capacity, voltage drop probability, and short-time frequency response capability; calculate the channel transmission capacity index by combining voltage level and feeder capacity; calculate the channel stability index by combining voltage drop probability and short-time frequency response capability; calculate the channel comprehensive index by combining the channel transmission capacity index and channel stability index; classify the channels in the distribution network data using the channel comprehensive index to obtain channel categories, namely: main channels and redundant channels; construct a distribution network channel topology map using the functional labels of the distribution network sub-areas and the channel categories.

[0111] Collect distribution network channel data, which includes voltage levels, feeder capacity, voltage drop probability, and short-time frequency impact capability of channels between distribution network sub-areas.

[0112] It should be noted that voltage level refers to the voltage level of each channel (e.g., 10kV, 35kV), which directly affects the power transmission capacity and applicable range. Feeder capacity indicates the maximum transmission capacity of the channel (e.g., current carrying capacity, power limit), used to assess whether it meets load demand or renewable energy access requirements. Voltage sag probability: the likelihood of voltage drop during a fault or sudden load change, which needs to be considered in conjunction with historical fault records or simulation data. Short-time frequency response capability: the channel's adjustment speed and stability during grid frequency fluctuations, which needs to be based on the dynamic response parameters of equipment (e.g., inverters, energy storage).

[0113] It should be noted that distribution network channel data refers to channel data connecting different distribution network sub-areas, excluding channel data within distribution network sub-areas, such as the connection channel data between distribution network sub-area A and distribution network sub-area B.

[0114] The channel transmission capacity index is calculated by combining the voltage level and feeder capacity:

[0115]

[0116] in, V represents the channel transmission capacity index, where V is the voltage level. Where F is the highest voltage of the channel and F is the feeder capacity. This represents the maximum feeder capacity of the channel.

[0117] The channel transmission capacity index is calculated by combining voltage level and feeder capacity. Voltage level represents the maximum voltage level the channel can handle for power transmission; higher voltage levels allow for greater power transmission and are suitable for long-distance, high-load power transmission. Feeder capacity reflects the channel's maximum transmission capacity under certain voltage conditions; it represents the maximum current or power the channel can carry. Feeder capacity is closely related to current and determines the effective power load the channel can handle without overload or faults. Combining these two factors provides a more comprehensive assessment of the channel's transmission capacity.

[0118] The channel stability index is calculated by combining the voltage drop probability and short-time frequency response capability:

[0119]

[0120] in, Here, D is the channel stability index, D is the voltage sag probability, and Q is the short-time frequency response capability. This represents the maximum probability of voltage drop. This represents the maximum short-time frequency response capability.

[0121] The channel stability index is calculated by combining voltage sag probability and short-time frequency response capability to assess the stability of distribution network channels under different operating conditions. Voltage sag probability reflects the frequency at which voltage may drop during channel operation. Voltage sags are typically caused by factors such as load fluctuations, equipment failures, or external interference. Frequent voltage sags can affect the normal operation of the distribution network and even trigger power equipment failures. Therefore, voltage sag probability is an important indicator for measuring the stability of distribution network channels. Short-time frequency response capability measures the channel's response capability to instantaneous load changes or power system fluctuations. Frequency response capability reflects the ability of a distribution network channel to quickly recover to a stable state when affected by sudden load fluctuations or the integration of new energy sources. Combining these two factors, the calculated channel stability index can comprehensively reflect the channel's stability under conditions such as voltage sags and load fluctuations. Voltage sag probability provides information on the potential risks of the channel in long-term operation, while short-time frequency response capability reflects the channel's recovery capability in the face of instantaneous changes.

[0122] The overall channel index is calculated by combining the channel transmission capacity index and the channel stability index:

[0123]

[0124] Wherein, ZN is the channel composite index. This is the channel transmission capacity index. This is the channel stability index.

[0125] A comprehensive channel index is calculated by combining the channel transmission capacity index and the channel stability index, comprehensively evaluating the overall performance of distribution network channels in terms of power transmission capacity and operational stability. The channel transmission capacity index mainly reflects the maximum power transmission capacity of the channel under different voltage levels and feeder capacities, while the channel stability index reflects the channel's ability to cope with unstable factors such as voltage dips and frequency fluctuations during operation. Combining these two indices allows for a comprehensive evaluation of the channel's overall performance from both capacity and stability perspectives, thus providing a more accurate basis for grid design and optimization. The calculation of the comprehensive channel index not only helps determine which channels are more suitable for handling critical loads or new energy access, but also ensures that the grid maintains high reliability and stability during emergencies or load fluctuations, improving the overall operational efficiency and resilience of the grid.

[0126] Main Road: Comprehensive Index of Throughpass With preset channel threshold Comparison, > The main trunk line is a channel with high voltage level, large feeder capacity, and low voltage drop probability, which is more suitable for carrying critical loads or power transmission to centralized access points of new energy sources.

[0127] Redundant Channels: Channel Composite Index Less than the preset channel threshold Channels with lower ratings are designated as redundant channels. These redundant channels are used for fault switching or load transfer to improve grid resilience. For example, in the event of sudden changes in photovoltaic output or equipment failure, quickly switching to a redundant channel can avoid the risk of power outages.

[0128] A distribution network channel topology map is constructed using the functional labels of the distribution network sub-areas and channel categories.

[0129] It should be noted that in the distribution network channel topology map, nodes represent sub-regions: Sub-region functional labels are marked with graphical nodes (e.g., color differentiation: red - daytime peak type, green - photovoltaic sensitive type, blue - industrial load type). Edges represent channels: directed edges are drawn based on the scoring results, with line width positively correlated with the score; dashed lines represent redundant channels. Dynamic weight labeling: key parameters are labeled next to the channel (e.g., "35kV, capacity 5MW, voltage drop probability 2%"). For example: Sub-region A (photovoltaic sensitive type): connects two channels, channel 1 (main channel, 35kV, capacity 10MW, score 0.85) and channel 2 (redundant channel, 10kV, capacity 3MW, score 0.6). Sub-region B (industrial load type): connected to sub-region A via channel 1, and connected to sub-region C via another redundant channel. Sub-region C (daytime peak type): connected to sub-region B via a redundant channel. The generated map shows a thick solid line (high-scoring main channel) between sub-regions A and B, labeled "35kV, 10MW". A thin dashed line (low-score redundant channel) is drawn between sub-regions B and C, labeled "10kV, 3MW, 5% probability of voltage drop". Sub-region A is labeled with the function tag "PV sensitive" and linked to the PV power station icon.

[0130] Step S5: Based on the distribution network channel topology map, obtain the number of main road connections and redundant channel connections of the distribution network sub-area; by combining the functional tags of the distribution network sub-area, the number of main road connections and redundant channel connections of the distribution network sub-area, calculate the energy storage index of the distribution network sub-area; deploy energy storage units for the distribution network sub-area based on the energy storage index of the distribution network sub-area.

[0131] Based on the distribution network channel topology diagram, the number of main channel connections and the number of redundant channel connections in the distribution network sub-area are obtained.

[0132] It should be noted that, based on the aforementioned distribution network channel topology map, each node in the distribution network channel topology map represents a distribution network sub-area, and each edge in the distribution network channel topology map represents a connection channel of the distribution network sub-area. Therefore, the distribution network channel topology map contains the functional label of each distribution network sub-area, the number of main road connections and the number of redundant channel connections of the distribution network sub-area.

[0133] Based on the aforementioned distribution network channel topology map, and by combining the functional tags of the distribution network sub-regions, the number of main channel connections and the number of redundant channel connections in the distribution network sub-regions, the energy storage index of the distribution network sub-regions is calculated:

[0134]

[0135] in, This represents the energy storage index of the i-th distribution network sub-region. This represents the number of main road connections in the i-th distribution network sub-area. The functional label represents the cluster to which the i-th distribution network sub-area belongs, and This represents the number of redundant channel connections in the i-th distribution network sub-area. and is a weighting coefficient used to smooth the importance of the main channel and the redundant channel, and a and b are sublinear exponents used to adjust the impact of the main channel and the redundant channel on the energy storage configuration.

[0136] It should be noted that the functional labels for distribution network sub-area clusters The numerical levels are defined as follows: 3 for photovoltaic sensitive type (high volatility risk), 2 for daytime peak type (medium volatility), and 1 for industrial load type (low volatility).

[0137] The distribution network sub-area energy storage index is used to divide the distribution network sub-area into central energy storage systems and edge energy storage units. If the energy storage index exceeds the preset threshold, a centralized energy storage system will be deployed in the distribution network sub-area; the energy storage index of the distribution network sub-area... If the value is less than the preset energy storage index threshold, then edge energy storage units will be deployed in the sub-area of ​​the distribution network.

[0138] Step S6: Collect adjustable load data between sub-sections of the distribution network, construct an energy storage load coordination model based on the adjustable load data between the sub-sections of the distribution network and the energy storage unit, and realize the power consumption planning of the distribution network for new energy access through the energy storage load coordination model.

[0139] To collect adjustable load data for sub-areas of the distribution network, it is necessary to identify the specific types of adjustable loads in the area (such as residential air conditioners, electric vehicle charging piles, industrial interruptible equipment, commercial lighting, etc.) and to collect their basic power consumption parameters, including rated power, typical power consumption periods, number of devices, and distribution density.

[0140] Construct an energy storage load coordination model, the objective function of which is:

[0141]

[0142] in, Let be the objective function. To optimize the total number of time periods within the time window, M represents the total number of energy storage devices, and M represents the total number of adjustable loads. For the i-th energy storage device during the time period Operating costs For adjustable load m during time period The power adjustment amount is positive for increasing the load and negative for reducing the load. The adjustment compensation unit price is the adjustable load m.

[0143] It should be noted that, To optimize the total number of time periods in the time window, let's define the time for the next T time periods. The adjustment compensation unit price for adjustable load m is set according to different adjustment priorities and compensation coefficients based on load type. For example, industrial interruptible equipment can be set with a high compensation unit price (because adjustment affects production); residential air conditioning can be set with a low compensation unit price (because adjustment is more flexible).

[0144] Calculate the operating cost of energy storage equipment based on the different types of energy storage units:

[0145]

[0146] in, For the i-th energy storage device during the time period Operating costs This represents the charging power of the i-th energy storage device during time period t. This represents the discharge power of the i-th energy storage device during time period t. , This represents the charging cost coefficient and discharging cost coefficient of a centralized energy storage system during time period t. , This represents the charging cost coefficient and discharging cost coefficient of the edge energy storage unit in time period t.

[0147] The constraints of the energy storage load coordination model are: , ,in, and This indicates the minimum and maximum allowable state of charge (SOC) of the energy storage device. The state of charge (SOC) of the i-th energy storage device during time period t represents the proportion of the energy currently stored by the device to its rated capacity, typically ranging from 0 to 1. Let m be the minimum power adjustment allowed for the adjustable load of type m during time period t. This represents the maximum allowable power adjustment for the m-th type of adjustable load during time period t. This represents the actual power adjustment of the m-th type of adjustable load during time period t.

[0148] Power constraints: ,in, The total power output of new energy sources during time period t. Let be the discharge power of the i-th energy storage device in time period t, that is, the power released by the energy storage to the grid. The rigid load power during time period t, This is the reference power for the m-th type of adjustable load. Let be the power adjustment amount of the m-th type of adjustable load during time period t. Let be the charging power of the i-th energy storage device during time period t, where i is the i-th energy storage device, n is the total number of energy storage devices, m is the m-th type of adjustable load, and M is the total number of adjustable loads.

[0149] The predictive data inputs for the energy storage load coordination model are: the total output power of new energy sources, the power of rigid loads and the baseline power of adjustable loads, and the charging and discharging cost coefficient of the energy storage station.

[0150] It should be noted that the total output power of new energy sources is predicted for future time period T by real-time monitoring of renewable energy sources such as photovoltaics and wind power, and meteorological data. For example, meteorological data such as wind speed, light intensity, and temperature obtained from meteorological monitoring platforms, as well as historical data on the total output power of new energy sources, are collected. Machine learning models, such as support vector machine regression models, are used to predict the total output power of new energy sources for future time period T. Furthermore, the load is divided into rigid load and adjustable load. Rigid load includes residential electricity consumption and infrastructure, while adjustable load includes air conditioning, electric vehicle charging stations, etc. The charging cost coefficient represents the unit cost of energy storage purchasing electricity from the grid for charging during a certain period, and is usually directly taken as the time-of-use electricity price for that period (e.g., off-peak electricity price of 0.4 yuan / kWh). The discharging cost coefficient reflects the revenue or cost of energy storage supplying power to the grid during discharging, and is usually set as a negative value or a discount value of the time-of-use electricity price (e.g., 80% of the peak electricity price of 1.2 yuan / kWh). For example, during peak hours, high electricity prices correspond to high charging cost coefficients, and the model will prioritize reducing charging and increasing discharging to lower the total cost; during off-peak hours, low electricity prices correspond to low charging cost coefficients, and the model tends to increase charging to store low-priced electricity. This design dynamically adjusts energy storage behavior through electricity price signals, responding to price fluctuations in the electricity market while optimizing the charging and discharging strategy of energy storage, ultimately minimizing the total operating cost.

[0151] Optimization solution of the energy storage load coordination model: Rolling time-domain optimization is adopted at each time point. The input data for the energy storage load coordination model includes: total output power of new energy sources, baseline power of rigid loads and adjustable loads, and charging / discharging cost coefficients of the energy storage station. Using optimization tools such as Gurobi and CPLEX, a mixed-integer linear programming problem is solved to minimize the objective function of the energy storage load coordination model while satisfying its constraints and power constraints. The final output yields the decision variables for the future time period T.

[0152]

[0153] in, For decision variables in the future time period T, This represents the charging power of the i-th energy storage device during time period t. This represents the discharge power of the i-th energy storage device during time period t. For adjustable load m during time period The power adjustment amount, where T is the future time period T.

[0154] It should be noted that the decision variables for the future T-period provide a dynamically optimized operating strategy for the distribution network to address the uncertainty of renewable energy output and load demand, while achieving a balance between economy and safety. During periods of low electricity prices, the energy storage load coordination model prioritizes charging energy storage devices to store more low-priced electricity, guiding some adjustable loads (such as electric vehicles and water heaters) to increase electricity consumption to absorb potentially excess renewable energy. During periods of high electricity prices, the energy storage load coordination model prioritizes discharging energy storage stations to provide higher-priced electricity (equivalent to electricity sales revenue), and prioritizes calling adjustable loads with low compensation unit prices (such as residential air conditioning group control for gentle temperature increases) to smooth out peak flows, minimizing or avoiding the use of loads with high compensation unit prices (such as industrial equipment interruptions affecting production). The energy storage load coordination model compares the charging and discharging cost differences between centralized energy storage systems and edge energy storage units, optimizing their calling order and extent. For example, it prioritizes using lower-cost edge energy storage units to smooth voltage fluctuations within sub-regions. During the predicted peak afternoon hours for renewable energy output, the energy storage load coordination model schedules energy storage charging in advance when electricity prices are lower at midday. This prepares for efficient absorption of surplus energy output in the afternoon and reduces peak electricity purchases. Conversely, during the predicted evening hours for renewable energy output, the model guides some adjustable loads (such as electric vehicles) to partially charge in advance when renewable energy output is sufficient and electricity prices are acceptable. This avoids concentrated high-power charging during the evening peak hours, which could exacerbate grid pressure.

[0155] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, the phrase "comprising an element defined as..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0156] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A distribution network planning method adapted to distributed renewable energy access, characterized in that: Includes the following steps: Step S1: Collect and preprocess the operation data of the distribution network sub-area to obtain the operation dataset of the distribution network sub-area. Calculate the entropy value of the operation dataset of the distribution network sub-area using the entropy weight method to obtain the feature weight of each feature column of the operation dataset of the distribution network sub-area. Step S2: Cluster the power distribution sub-area operation dataset using the KMEANS algorithm to obtain power distribution sub-area clusters; calculate the weighted cluster center by combining the cluster center of the power distribution sub-area clusters with the feature weight of each feature column of the power distribution sub-area operation dataset; Step S3: Calculate the scores of the weighted cluster centers using a sorting method that approximates the ideal solution to obtain the maximum approximation score of the weighted cluster centers. Then, use the maximum approximation score to divide the distribution network sub-areas into functional labels to obtain the functional labels of the distribution network sub-areas. Step S4: Collect distribution network channel data, including voltage level, feeder capacity, voltage drop probability, and short-time frequency response capability; calculate the channel transmission capacity index by combining voltage level and feeder capacity; calculate the channel stability index by combining voltage drop probability and short-time frequency response capability; calculate the channel comprehensive index by combining the channel transmission capacity index and channel stability index; classify the channels in the distribution network data using the channel comprehensive index to obtain channel categories, namely: trunk channels and redundant channels; construct a distribution network channel topology map using the functional labels of the distribution network sub-areas and the channel categories. Step S5: Based on the distribution network channel topology map, obtain the number of main road connections and redundant channel connections of the distribution network sub-area; by combining the functional tags of the distribution network sub-area, the number of main road connections and redundant channel connections of the distribution network sub-area, calculate the energy storage index of the distribution network sub-area; deploy energy storage units for the distribution network sub-area based on the energy storage index of the distribution network sub-area. Step S6: Collect adjustable load data between sub-sections of the distribution network, construct an energy storage load coordination model based on the adjustable load data between the sub-sections of the distribution network and the energy storage unit, and realize the power consumption planning of the distribution network for new energy access through the energy storage load coordination model.

2. The distribution network planning method adapted to distributed new energy access according to claim 1, characterized in that: The process of calculating the entropy value of the distribution network operation dataset using the entropy weight method to obtain the feature weight of each feature column of the distribution network operation dataset includes the following specific steps: The entropy value of the characteristic column in the sub-area operation data of the distribution network is calculated using the entropy weight method: ； in, is the entropy value of the j-th feature in the distribution network sub-area operation data, and n is the number of samples in the distribution network sub-area operation data. This refers to the proportion of data from the i-th distribution network sub-area in the j-th feature column, where j represents the index of the j-th feature and i represents the data from the i-th distribution network sub-area. Calculate the weight of each feature in the sub-area operation data of the distribution network: ； in, is the weight of the j-th feature of the distribution network sub-area operation data, where J is the total number of features of the distribution network sub-area operation data. It is the entropy value of the j-th feature of the distribution network sub-area operation data, where j represents the index of the j-th feature.

3. The distribution network planning method adapted to distributed new energy access according to claim 2, characterized in that: The step of calculating the weighted cluster center by combining the cluster centers of the distribution network sub-area clusters and the feature weights of each feature column in the distribution network sub-area operation dataset includes the following steps: By combining the cluster centers of the distribution network sub-area clusters with the aforementioned feature weights, the weighted cluster centers are calculated: ； in, This represents the value of the k-th cluster center after weighting the j-th feature. This represents the feature weight of the j-th feature. Indicates the first sub-region cluster of the distribution network The value of the cluster center on the j-th feature, where j represents the index of the j-th feature, and K represents the total number of cluster centers.

4. The distribution network planning method adapted to distributed new energy access according to claim 3, characterized in that: The step of calculating the scores of the weighted cluster centers using a sorting method that approximates the ideal solution to obtain the maximum approximation score of the weighted cluster centers includes the following steps: For a distribution network sub-area with function label L, its positive ideal solution is: The negative ideal solution is ; For each distribution network sub-area cluster, calculate the distance between its weighted cluster center and the positive ideal solution for different functional labels of the distribution network sub-area: ； in, The distance between the k-th distribution network sub-area cluster and the positive ideal solution of the L-th functional label, where j represents the index of the j-th feature, and J is the total number of features in the distribution network sub-area operational data. This represents the value of the k-th cluster center after weighting the j-th feature. This represents the value of the positive ideal solution of the Lth functional label in the jth feature; For each distribution network sub-region cluster, calculate the distance between its weighted cluster center and the negative ideal solution of different functional labels of the sub-region: ； in, The distance between the k-th distribution network sub-area cluster and the negative ideal solution of the L-th functional label, where j represents the index of the j-th feature, and J is the total number of features in the distribution network sub-area operational data. This represents the value of the k-th cluster center after weighting the j-th feature. This represents the value of the negative ideal solution of the Lth functional label in the jth feature; For each distribution network sub-area cluster, the maximum approximation score of the cluster is calculated by using the distances between its weighted cluster center and the positive ideal solutions of different functional labels of the sub-area. ； in, Indicates the first The maximum approximation score of each cluster reflects how close it is to the ideal solution. The distance between the k-th distribution network sub-area cluster and the positive ideal solution of the L-th functional label. The distance between the k-th distribution network sub-area cluster and the negative ideal solution of the L-th functional label is given by max(), which is a maximum value function representing the maximum approximation score between the k-th distribution network sub-area cluster and the L functional labels. The approximation score between the k-th distribution network sub-area cluster and the L-th functional label.

5. A distribution network planning method adapted to distributed renewable energy access according to claim 4, characterized in that: The process of dividing the distribution network sub-regions into functional labels based on the maximum approximation score to obtain the functional labels of the distribution network sub-regions includes the following specific steps: For the k-th distribution network sub-area cluster, by calculating the approximation score between its weighted cluster center and different functional labels, the maximum approximation score between the k-th distribution network sub-area cluster and the L-th functional label is taken as the maximum approximation score of the k-th distribution network sub-area cluster. Therefore, the functional label of the k-th distribution network sub-area cluster is... .

6. A distribution network planning method adapted to distributed new energy access according to claim 5, characterized in that: The calculation of the channel transmission capacity index by combining voltage level and feeder capacity includes the following specific steps: The channel transmission capacity index is calculated by combining the voltage level and feeder capacity: ； in, V represents the channel transmission capacity index, where V is the voltage level. Where F is the highest voltage of the channel and F is the feeder capacity. This represents the maximum feeder capacity of the channel.

7. A distribution network planning method adapted to distributed new energy access according to claim 6, characterized in that: The channel stability index is calculated by combining voltage drop probability and short-time frequency response capability, including the following specific steps: The channel stability index is calculated by combining the voltage drop probability and short-time frequency response capability: ； in, Here, D is the channel stability index, D is the voltage sag probability, and Q is the short-time frequency response capability. This represents the maximum probability of voltage drop. This represents the maximum short-time frequency response capability.

8. A distribution network planning method adapted to distributed renewable energy access according to claim 7, characterized in that: The process of calculating the comprehensive channel index by combining the channel transmission capacity index and the channel stability index includes the following specific steps: The overall channel index is calculated by combining the channel transmission capacity index and the channel stability index: ； Wherein, ZN is the channel composite index. This is the channel transmission capacity index. This is the channel stability index.

9. A distribution network planning method adapted to distributed new energy access according to claim 8, characterized in that: The calculation of the energy storage index of a distribution network sub-region by combining the functional tags of the sub-region, the number of main network connections and the number of redundant channel connections in the sub-region, includes the following specific steps: Based on the aforementioned distribution network channel topology map, and by combining the functional tags of the distribution network sub-regions, the number of main channel connections and the number of redundant channel connections in the distribution network sub-regions, the energy storage index of the distribution network sub-regions is calculated: ； in, This represents the energy storage index of the i-th distribution network sub-region. This represents the number of main road connections in the i-th distribution network sub-area. The functional label represents the cluster to which the i-th distribution network sub-area belongs, and This represents the number of redundant channel connections in the i-th distribution network sub-area. and is a weighting coefficient used to smooth the importance of the main channel and the redundant channel, and a and b are sublinear exponents used to adjust the impact of the main channel and the redundant channel on the energy storage configuration.

10. A distribution network planning method adapted to distributed new energy access according to claim 9, characterized in that: The construction of the energy storage load coordination model includes the following specific steps: Construct an energy storage load coordination model, the objective function of which is: ； in, Let be the objective function. To optimize the total number of time periods within the time window, n is the number of samples of operational data for the distribution network sub-area, and M is the total number of adjustable loads. For the i-th energy storage device during the time period Operating costs are related to time-of-use electricity pricing. For adjustable load m during time period The power adjustment amount is positive for increasing the load and negative for reducing the load. The adjustment compensation unit price for adjustable load m; Calculate the operating cost of energy storage equipment based on the different types of energy storage units: ； in, For the i-th energy storage device during the time period Operating costs This represents the charging power of the i-th energy storage device during time period t. This represents the discharge power of the i-th energy storage device during time period t. , This represents the charging cost coefficient and discharging cost coefficient of a centralized energy storage system during time period t. , This represents the charging cost coefficient and discharging cost coefficient of the edge energy storage unit in time period t; The constraints of the energy storage load coordination model are: , , ,in, and This indicates the minimum and maximum allowable state of charge (SOC) of the energy storage device. Let represent the state of charge (SOC) of the i-th energy storage device during time period t, indicating the proportion of the energy currently stored by the device to its rated capacity, ranging from 0 to 1. Let m be the minimum power adjustment allowed for the adjustable load of type m during time period t. This represents the maximum allowable power adjustment for the m-th type of adjustable load during time period t. This represents the actual power adjustment of the m-th type of adjustable load during time period t. Power constraints: ,in, The total power output of new energy sources during time period t. Let be the discharge power of the i-th energy storage device in time period t, that is, the power released by the energy storage to the grid. The rigid load power during time period t, This is the reference power for the m-th type of adjustable load. Let be the power adjustment amount of the m-th type of adjustable load during time period t. Let represent the charging power of the i-th energy storage device during time period t, where i is the i-th energy storage device, n is the number of samples of the distribution network sub-area operation data, m is the m-th type of adjustable load, and M is the total number of adjustable loads.

Images