A power distribution network evolution form division method considering voltage out-of-limit and harmonic over-limit

Abstract

The application discloses a power distribution network evolution mode division method considering voltage out-of-limit and harmonic over-limit, and belongs to the technical field of intelligent power grid operation management and control, and comprises the following steps: constructing a multi-resource grid connection scene of multiple flexible resources of a power distribution network; performing power flow calculation and harmonic power flow calculation on the multi-resource grid connection scene, acquiring voltage amplitude, voltage harmonic content and voltage total harmonic distortion rate of each node at different time sections, and constructing a power distribution network voltage out-of-limit matrix and a power distribution network harmonic over-limit matrix; based on the power distribution network voltage out-of-limit matrix, performing primary division on the multi-resource grid connection scene by using a density clustering DBSCAN algorithm, and based on the primary division result, performing secondary division on the multi-resource grid connection scene by using the DBSCAN algorithm according to the power distribution network harmonic over-limit matrix. The method fully reflects the complexity and quality constraint of the operation state in the mode division process, and the result is more comprehensive and more in line with the actual operation characteristics.

Description

Technical Field

[0001] This application relates to the field of smart grid operation and control technology, and in particular to a method for classifying distribution network evolution patterns that considers voltage overruns and harmonic exceedances. Background Technology

[0002] With the rapid development of new power systems, flexible power sources such as distributed photovoltaic power generation, user-side energy storage systems, and electric vehicle charging piles are being connected to the distribution network in large numbers, significantly improving the system's flexibility and its capacity to accommodate renewable energy. However, while these new resources improve energy efficiency, they also pose new challenges to the operational stability and power quality management of the distribution network, especially with problems such as node voltage exceeding limits and harmonic exceedances becoming more frequent and complex.

[0003] The current power distribution network is evolving from a traditional static structure into a complex system with dynamic evolutionary characteristics. Its operating state constantly changes with the scale of resource integration, time-of-day load characteristics, and control strategies, resulting in diverse operating modes. However, most existing research focuses only on voltage levels or harmonic indicators under static or specific scenarios, lacking system modeling methods oriented towards the evolutionary process. This makes it difficult to comprehensively characterize the diversity and evolutionary patterns of the power distribution network's operating state under different resource penetration levels and integration structures. Furthermore, existing methods often fail to fully consider the stochastic characteristics of flexible resource integration and the spatiotemporal non-stationarity of power quality issues. They lack the ability to co-model voltage exceedance and harmonic exceedance problems, and even more so, they lack data-driven partitioning methods based on actual physical characteristics. Given the limited resources for power quality management, accurately delineating the evolutionary form of the power distribution network and identifying the key management priorities and planning directions at each stage has become a crucial prerequisite for power distribution system operation analysis and control optimization.

[0004] Therefore, there is an urgent need to propose a morphological classification method that integrates the characteristics of multiple types of flexible resource access and the results of voltage and harmonic collaborative assessment, and can systematically and accurately reflect the differences in the evolution trend and governance needs of the distribution network operation, so as to support key decision-making tasks such as power quality improvement, resource optimization and allocation, and future distribution system evolution path planning. Summary of the Invention

[0005] In view of the above-mentioned shortcomings in the prior art, this application provides a method for classifying the evolution of distribution network considering voltage overruns and harmonic exceedances, so as to solve the problems in the background art.

[0006] To achieve the aforementioned objectives, the technical solution adopted in this application is as follows:

[0007] First aspect:

[0008] This application provides a method for classifying distribution network evolution patterns considering voltage over-limit and harmonic exceedance, including:

[0009] Construct multi-resource grid connection scenarios for various flexible resources in the power distribution network;

[0010] Power flow and harmonic power flow calculations are performed for multi-resource grid-connected scenarios to obtain the voltage amplitude, voltage harmonic content and total harmonic distortion rate of each node at different time sections, and to construct the distribution network voltage over-limit matrix and the distribution network harmonic exceedance matrix.

[0011] Based on the voltage limit exceedance matrix of the distribution network, the density clustering DBSCAN algorithm is used to perform a first division of the multi-resource grid-connected scenario. Based on the first division result, the DBSCAN algorithm is used to perform a second division of the multi-resource grid-connected scenario according to the harmonic exceedance matrix of the distribution network.

[0012] Furthermore, the construction of a multi-resource grid connection scenario with multiple flexible resources in the distribution network includes:

[0013] Construct a vector combining the penetration rates of various flexible resources in the power distribution network;

[0014] Calculate the node connection probability of various flexible resources in the distribution network at different nodes;

[0015] Based on the penetration rate of various flexible resources in the combination vector of multiple flexible resources in the distribution network and the node grid connection probability of various flexible resources at different nodes, a single resource grid connection scenario is constructed.

[0016] By merging single-resource grid connection scenarios according to resource grid connection nodes, we obtain multi-resource grid connection scenarios.

[0017] Furthermore, the calculation of the node connection probability of various flexible resources in the distribution network at different nodes includes:

[0018] Considering the topology and electrical properties of nodes in a distribution network, calculate the node importance coefficient. PR value;

[0019] Based on node importance coefficient PR The value is used to calculate the probability of multiple types of flexible resources connecting to the network at different nodes.

[0020] Furthermore, the construction of a single-resource grid connection scenario based on the penetration rate combination vector of various flexible resources in the distribution network and the node grid connection probability of various flexible resources at different nodes includes:

[0021] Random selection Each node is used as the initial node in a single resource grid connection scenario, and the set probability of the initial node is calculated.

[0022] Randomly select several nodes, change the access status of the resources corresponding to the selected nodes, and obtain the probability of the node set of new candidate scenarios;

[0023] Based on the initial node set probability and the new candidate scene node set probability, the Markov chain state transition method is used to calculate the acceptance probability of accepting the newly generated candidate scene and generate a random number.

[0024] Based on the obtained acceptance probability and the generated random number, determine whether to accept the newly generated candidate scene;

[0025] Based on the judgment results, a single resource grid connection scenario is obtained.

[0026] Furthermore, the step of determining whether to accept the newly generated candidate scene based on the obtained acceptance probability and the generated random number includes:

[0027] If the random number is less than the acceptance probability, the state transitions, the candidate scenario is accepted as the new grid connection scenario, and the new grid connection scenario is used as the starting point for the next iteration; otherwise, the current scenario is retained.

[0028] Furthermore, the construction of the distribution network voltage over-limit matrix includes:

[0029] Calculate the voltage deviation amplitude based on the voltage amplitude of each node at different time sections and the nominal voltage of the distribution network system;

[0030] The voltage over-limit judgment coefficient is obtained based on the voltage deviation magnitude and the allowable voltage deviation of the distribution network;

[0031] Based on the voltage deviation amplitude and the voltage over-limit judgment coefficient, a voltage over-limit matrix reflecting the distribution network under different time sections is constructed.

[0032] Furthermore, the construction of the distribution network harmonic exceedance matrix includes:

[0033] Calculate the degree of voltage total harmonic distortion (THD) exceeding the standard and the judgment coefficient of the degree of voltage THD exceeding the standard, and construct the system voltage THD characteristic matrix;

[0034] Obtain the voltage content of each harmonic of the system and construct the voltage content matrix of each harmonic of the system;

[0035] Based on the system voltage total harmonic distortion rate characteristic matrix and the system harmonic voltage content matrix, a matrix reflecting the harmonic exceedance of the distribution network under different time sections is constructed.

[0036] Furthermore, the step of using the density clustering DBSCAN algorithm to partition the multi-resource grid-connected scenario based on the distribution network voltage over-limit matrix includes:

[0037] Based on the Hausdorff distance method, the similarity of voltage over-limit features of the scene is calculated;

[0038] Set the neighborhood radius of the scene voltage similarity and the sample radius of the distribution network voltage over-limit matrix of the scene as a threshold of the number of voltage over-limit matrix samples of other scenes in the neighborhood of the scene voltage similarity;

[0039] Randomly select a scene and calculate all scenes whose voltage over-limit feature similarity to the selected scene is less than or equal to the scene voltage similarity neighborhood radius;

[0040] If the total number of all obtained scenarios is less than the threshold, the selected scenario is discarded, a new scenario is selected, and all scenarios with voltage limit violation feature similarity to the selected scenario less than or equal to the voltage similarity neighborhood radius of the scenario are calculated. The process continues to determine whether the total number of all obtained scenarios is less than the threshold. Otherwise, the selected scenario is taken as the core scenario, and other scenarios are determined based on the core scenario.

[0041] If it is a core scenario, the core scenarios are grouped into a cluster; otherwise, it is determined whether the unselected scenarios are core scenarios, and the core scenarios are grouped into a cluster, until all scenarios are divided into different clusters according to the distribution network voltage over-limit matrix, thus completing one division of the distribution network evolution form.

[0042] Furthermore, based on the initial partitioning result, the DBSCAN algorithm is used to perform a secondary partitioning of the multi-resource grid-connected scenario according to the distribution network harmonic exceedance matrix, including:

[0043] In each cluster of the initial division, the harmonic similarity of the scenarios is calculated, and the DBSCAN algorithm is used to re-cluster all multi-resource grid-connected scenarios to achieve a secondary division of the distribution network evolution pattern.

[0044] The second aspect:

[0045] This application provides a distribution network evolution pattern classification device that considers voltage over-limit and harmonic exceedance, including:

[0046] The grid connection scenario generation module is used to generate multi-resource grid connection scenarios for various flexible resources in the distribution network;

[0047] The calculation module is used to perform power flow calculation and harmonic power flow calculation in multi-resource grid-connected scenarios, and to obtain the voltage amplitude, voltage harmonic content and total harmonic distortion rate of each node at different time sections.

[0048] The power quality exceedance matrix construction module is used to construct the distribution network voltage exceedance matrix and the distribution network harmonic exceedance matrix based on the voltage amplitude, voltage harmonic content and total harmonic distortion rate of each node at different time sections.

[0049] The grid connection scenario segmentation module is used to perform two-stage segmentation of multi-resource grid connection scenarios based on the distribution network voltage over-limit matrix and the distribution network harmonic over-limit matrix.

[0050] The beneficial effects of this application are:

[0051] This application provides a method for classifying the evolutionary form of distribution networks considering voltage exceedance and harmonic distortion. By constructing a multi-dimensional power quality index system for distribution network nodes, it collaboratively analyzes the voltage exceedance amplitude and harmonic distortion rate of nodes. The method fully reflects the complexity of the operating state and quality constraints in the form classification process, resulting in more comprehensive results that better match actual operating characteristics. Furthermore, it combines the density clustering algorithm DBSCAN to classify the evolutionary form of distribution networks involving two power quality issues: voltage exceedance and harmonic distortion. This provides a reference for grid connection planning of photovoltaic, energy storage, and charging resources in different development stages of distribution networks. Attached Figure Description

[0052] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other embodiments can be obtained based on these drawings.

[0053] Figure 1 This is a flowchart illustrating a method for classifying the evolution of distribution networks considering voltage overruns and harmonic exceedances, as provided in an embodiment of this application. Detailed Implementation

[0054] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art based on this application are within the scope of protection of this application.

[0055] Example 1:

[0056] This application provides a method for classifying distribution network evolution patterns considering voltage over-limit and harmonic exceedance. This method can be found in [reference needed]. Figure 1 , Figure 1 The diagram shown is a flowchart illustrating a distribution network evolution morphology classification method considering voltage overruns and harmonic exceedances provided in this application, comprising:

[0057] S1: Construct a multi-resource grid connection scenario for various flexible resources in the distribution network.

[0058] S101: Construct a vector combining the penetration rates of various flexible resources in the distribution network.

[0059] In one embodiment of this application, the penetration rate of flexible resources such as distributed photovoltaics, energy storage, and electric vehicle charging piles follows a continuous uniform distribution as an example, with each flexible resource having a penetration rate range of... The sampling probabilities in Neamont-Carlo are equal. , These represent the maximum and minimum penetration rates of distributed photovoltaic power, energy storage, and electric vehicle charging stations, respectively. The maximum carrying capacity can be calculated based on the current technical standards for grid connection of different flexible resources, thereby obtaining the combination of penetration rates of various flexible resources in the distribution network. :

[0060]

[0061] in, , , These represent the penetration rates of distributed photovoltaic power, energy storage, and electric vehicle charging stations, respectively.

[0062] S102: Calculate the node connection probability of various flexible resources in the distribution network at different nodes.

[0063] In one embodiment of this application, due to the complexity of the distribution network topology and electrical characteristics, each node has different importance to the system operation, leading to a bias in the selection of nodes for flexible resources such as photovoltaics, energy storage, and electric vehicle charging piles for grid connection. This application uses the node importance coefficient as the probability of grid connection for multiple types of resources at each node in the distribution network. The PageRank algorithm was originally developed by calculating the PageRank of web pages. PR The ranking algorithm used to measure the importance of web pages on the internet can be similarly applied to assessing the importance of nodes in a power distribution network. Taking into account both the topological and electrical characteristics of the nodes, a node importance coefficient is calculated. PR value:

[0064]

[0065]

[0066] in, For the first Record each node after the next iteration. PR Value column vector, setting the initial values ​​of each node. PR The value is 1 for all nodes, until each node PR Stop iterating after the value converges, and output the value. , The transition matrix is ​​determined by the distribution network topology and electrical characteristics. It is the first This iteration covers all nodes. PR value, For the first Next iteration node of pr value, The number of distribution network nodes. This is a transpose.

[0067] According to each node of the distribution network PR The values ​​are used to calculate the connection probabilities of photovoltaic (PV), energy storage, and electric vehicle (EV) charging piles at different nodes. Distributed PV and energy storage, as power sources, are more important to the distribution network than EV charging piles as loads. However, compared to distributed PV, energy storage can stabilize the fluctuations in PV output. Therefore, the grid connection probability of energy storage at a node is related to... PR The value is positively correlated, making it more likely that energy storage can be connected to important nodes in the distribution network; the probability of charging piles connecting to the grid at nodes is positively correlated with... PR The values ​​are negatively correlated, reducing the likelihood of them being connected to important nodes in the distribution network. The calculation method for the grid connection probability of energy storage and electric vehicle charging pile resource nodes is as follows:

[0068]

[0069] in, and These represent the node grid connection probabilities for energy storage and electric vehicle charging piles, respectively. For nodes of PR The value, with the superscript -1 indicating the inverse matrix.

[0070] The nodes with a higher probability of grid connection for photovoltaic power are PR At nodes with appropriate values, the probability of grid connection of photovoltaic power at the node is related to... PR The relationship between the values ​​resembles the shape of a normal distribution or a bell curve. Therefore, the Gaussian function can be used to calculate the grid connection probability of photovoltaic nodes, as shown in the following formula:

[0071]

[0072] in, This represents the probability of a node being connected to the grid in a distributed photovoltaic system. The center of the probability distribution of photovoltaic nodes connected to the grid is determined by the number of distribution network nodes. Decide, The standard deviation represents the width of the probability distribution, which is determined by the generation requirements of the photovoltaic grid-connected scenario.

[0073] S103: Based on the penetration rate of various flexible resources in the combination vector of multiple flexible resource penetration rates in the distribution network and the node grid connection probability of various flexible resources at different nodes, construct a single resource grid connection scenario.

[0074] In one embodiment of this application, random selection k Each node is used as the initial node for single-resource network connection, and the probability of the node set is calculated. , The calculation formula is shown below:

[0075]

[0076] in, The set probability of the initial nodes. Values , and , indicating the first in the current scenario The probability of a node connecting to the network for a certain flexible resource.

[0077] Then, randomly select several nodes, change their resource access status, and obtain a new set of candidate scenario nodes with probabilities. .

[0078] Secondly, determine whether to accept the newly generated candidate scenario. Calculate the acceptance probability using the Markov chain state transition method. And generate a random number in the range (0,1). ,like If the current scenario is accepted, the state transitions, and the candidate scenario is accepted as the new grid connection scenario, serving as the starting point for the next iteration; otherwise, the current scenario is retained. The calculation formula is shown below:

[0079]

[0080] Repeat this step to obtain single-resource multi-grid-connected scenarios for photovoltaic, energy storage, and electric vehicle charging piles, respectively. The generation processes of the three single-resource grid-connected scenarios are independent of each other and do not affect each other.

[0081] S104: Combine the three grid connection scenarios into one according to the resource grid connection nodes to obtain a multi-resource grid connection scenario.

[0082] It is understandable that each multi-resource grid connection scenario corresponds to a certain penetration rate combination.

[0083] S2: Perform power flow calculation and harmonic power flow calculation for multi-resource grid-connected scenarios, obtain the voltage amplitude, voltage harmonic content and total harmonic distortion rate of each node under different time sections, and construct the distribution network voltage over-limit matrix and the distribution network harmonic over-limit matrix.

[0084] In one embodiment of this application, a matrix is ​​constructed to reflect voltage overruns in the distribution network at different time sections. :

[0085]

[0086] in, For the first In the scenario, the first Each node The voltage offset at any given time is calculated using the following formula:

[0087]

[0088] in, For the first In the scenario, the first Each node Voltage amplitude at time 10:00 The nominal voltage of the distribution network system. Number of scenes This is the voltage over-limit judgment coefficient, used to determine the scenario. Next node exist The rules for determining whether a voltage limit violation occurs at any given moment are as follows:

[0089]

[0090] in, This is the allowable voltage deviation for the distribution network. This value varies at different voltage levels and can be obtained through the international standard IEC 60038. It can reflect the frequency and extent of voltage over-limit occurrences at different time points and in the system under various scenarios. Characterizes the nodes exist A voltage exceeding the limit can occur at any time. If the value is greater than 0, it indicates the voltage has exceeded the upper limit; if the value is less than 0, it indicates the voltage has exceeded the lower limit. The larger the value, the more severe the voltage overshoot; if it equals 0, no voltage overshoot has occurred.

[0091] In one embodiment of this application, according to the international standard IEC 61000, a harmonic exceedance matrix of the distribution network under different time sections is constructed using the total harmonic distortion rate and the content of each harmonic voltage. :

[0092]

[0093] Among them, the block matrix This is the characteristic matrix of the total harmonic distortion rate of the system voltage. This is the matrix of voltage content for each harmonic of the system. It can be represented as:

[0094]

[0095] in, For the scene Next node exist The degree to which the total harmonic distortion (THD) of the voltage exceeds the limit at any given time can be calculated using the following formula:

[0096]

[0097] in, For the scene Next node exist The total harmonic distortion of voltage at time t. This is the limit for total harmonic distortion (THD) of voltage. The requirements for this value vary for different voltage levels and can be obtained from the international standard IEC 61000. This is a judgment coefficient for the degree to which the total harmonic distortion (THD) of the voltage exceeds the standard, used to determine the scenario. Next node exist The following rules govern whether the total harmonic distortion (THD) of the voltage exceeds the standard at any given time:

[0098]

[0099] Elements in the total distortion characteristic matrix of node voltage Represents a node exist The value is the difference between the total harmonic distortion rate of the actual voltage and the limit if harmonic distortion occurs; otherwise, it is 0.

[0100] The voltage content matrix of each harmonic is shown in the following formula:

[0101]

[0102] in, Representing a scene Middle node exist Moment Subharmonic voltage content.

[0103] Distribution network harmonic exceedance matrix Due to the large dimension and computational complexity, principal component analysis is used to... Dimensionality reduction is performed.

[0104] S3: Based on the voltage over-limit matrix of the distribution network, the density clustering DBSCAN algorithm is used to divide the multi-resource grid-connected scenario into two parts. Based on the result of the first part, the DBSCAN algorithm is used to divide the multi-resource grid-connected scenario into two parts according to the harmonic over-limit matrix of the distribution network.

[0105] In one embodiment of this application, taking the voltage over-limit problem as the object, the density clustering DBSCAN algorithm is used to divide all generated scenarios into two parts. Then, taking the harmonic over-limit problem as the object, the multiple scenarios of the distribution network are divided into two parts.

[0106] The specific steps for classifying the evolution of distribution networks considering power quality are as follows.

[0107] (1) Calculate the similarity of voltage limit exceedance features. This application defines a method for calculating scene voltage similarity based on Hausdorff distance. Hausdorff distance is a measure of the similarity between two datasets; the smaller the distance, the higher the similarity. The formula for calculating Hausdorff distance is:

[0108]

[0109] in, and These represent the one-way Hausdorff distances from scene E / F to scene F / E, respectively. and These represent the data points in the distribution network voltage over-limit matrix for scenarios E and F, respectively. The Euclidean distance between the two points is... The bidirectional Hausdorff distance between scenario E and scenario F reflects the maximum mismatch in voltage over-limit levels between the two operating scenarios. The smaller the distance, the lower the maximum mismatch and the higher the similarity of voltage over-limit features between the two scenarios.

[0110] (2) Setting parameters and Minpts . Represents the neighborhood radius of scene voltage similarity. Minpts The sample radius of the voltage over-limit matrix for this scenario is... The threshold for the number of voltage over-limit matrix samples in other scenarios within the domain.

[0111] (3) Scene division in one step. Randomly select a scene. By calculation, we find voltage over-limit characteristics with a similarity of less than or equal to that of the scenario. All scenarios: If the number of scenarios is less than Minpts Then the scene Abandoned; if the number of scenes is not less than Minpts ,but The core scene is defined; then other scenes are determined within the scene. of If a scenario is a core scenario within the neighborhood, then these scenarios belong to a cluster S. The remaining scenarios that were not selected are then evaluated until all scenarios have been evaluated and clustering is completed. All scenarios are then divided into different clusters based on the voltage limit matrix, thus achieving a classification of the evolution of the distribution network.

[0112] (4) Secondary scene segmentation. In each cluster of the first segmentation, the scene harmonic similarity is calculated, and the DBSCAN algorithm is used to repeat step (3) to cluster the scenes again, so as to realize the secondary segmentation of the power distribution network evolution form.

[0113] Example 2:

[0114] This application provides a distribution network evolution pattern classification device that considers voltage over-limit and harmonic exceedance, including:

[0115] The grid connection scenario generation module is used to generate multi-resource grid connection scenarios for various flexible resources in the distribution network;

[0116] The calculation module is used to perform power flow calculation and harmonic power flow calculation in multi-resource grid-connected scenarios, and to obtain the voltage amplitude, voltage harmonic content and total harmonic distortion rate of each node at different time sections.

[0117] The power quality exceedance matrix construction module is used to construct the distribution network voltage exceedance matrix and the distribution network harmonic exceedance matrix based on the voltage amplitude, voltage harmonic content and total harmonic distortion rate of each node at different time sections.

[0118] The grid connection scenario segmentation module is used to perform two-stage segmentation of multi-resource grid connection scenarios based on the distribution network voltage over-limit matrix and the distribution network harmonic over-limit matrix.

[0119] Example 3:

[0120] This application provides an electronic device, including:

[0121] At least one processor;

[0122] The memory is communicatively connected to the processor;

[0123] The memory stores instructions that can be executed by the at least one processor, which enable the at least one processor to perform any of the above-mentioned steps for classifying the distribution network evolution morphology considering voltage overruns and harmonic exceedances.

[0124] Example 4:

[0125] This application provides a computer-readable storage medium storing computer instructions that, when executed by a computer, implement any of the steps described above for classifying the distribution network evolution morphology considering voltage overruns and harmonic exceedances.

[0126] This application provides a method for classifying the evolutionary morphology of distribution networks considering voltage exceedance and harmonic distortion. By constructing a multi-dimensional power quality index system for distribution network nodes and collaboratively analyzing the voltage exceedance amplitude and harmonic distortion rate, the method fully reflects the complexity of operating states and quality constraints during the morphology classification process, resulting in more comprehensive results that better reflect actual operating characteristics. Furthermore, it employs density clustering to perform unsupervised learning on multi-dimensional feature data, eliminating the need for manually setting the number of categories. This automatically identifies the evolutionary patterns of the distribution network under different resource access conditions, effectively avoiding problems such as unreasonable manual threshold setting or weak identification capabilities. This improves the objectivity of morphology classification and provides a reference for grid connection planning of photovoltaic, energy storage, and charging resources in different development stages of the distribution network.

[0127] It should be noted that those skilled in the art will recognize that the embodiments described herein are for the purpose of helping readers understand the principles of this application, and should be understood as not limiting the scope of protection of this application to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations based on the technical teachings disclosed in this application without departing from the essence of this application, and these modifications and combinations are still within the scope of protection of this application.

Claims

1. A method for classifying distribution network evolution patterns considering voltage overruns and harmonic exceedances, characterized in that, include: Construct multi-resource grid connection scenarios for various flexible resources in the power distribution network; Power flow and harmonic power flow calculations are performed for multi-resource grid-connected scenarios to obtain the voltage amplitude, voltage harmonic content and total harmonic distortion rate of each node at different time sections, and to construct the distribution network voltage over-limit matrix and the distribution network harmonic exceedance matrix. Based on the voltage limit exceedance matrix of the distribution network, the density clustering DBSCAN algorithm is used to perform a first division of the multi-resource grid-connected scenario. Based on the first division result, the DBSCAN algorithm is used to perform a second division of the multi-resource grid-connected scenario according to the harmonic exceedance matrix of the distribution network.

2. The method for classifying distribution network evolution patterns considering voltage overruns and harmonic exceedances according to claim 1, characterized in that, The multi-resource grid connection scenario for constructing multiple flexible resources in the distribution network includes: Construct a vector combining the penetration rates of various flexible resources in the power distribution network; Calculate the node connection probability of various flexible resources in the distribution network at different nodes; Based on the penetration rate of various flexible resources in the combination vector of multiple flexible resources in the distribution network and the node grid connection probability of various flexible resources at different nodes, a single resource grid connection scenario is constructed. By merging single-resource grid connection scenarios according to resource grid connection nodes, we obtain multi-resource grid connection scenarios.

3. The method for classifying distribution network evolution patterns considering voltage overruns and harmonic exceedances according to claim 2, characterized in that, The calculation of the node connection probability of various flexible resources in the distribution network at different nodes includes: Considering the topology and electrical properties of nodes in a distribution network, calculate the node importance coefficient. PR value; Based on node importance coefficient PR The value is used to calculate the probability of multiple types of flexible resources connecting to the network at different nodes.

4. The method for classifying distribution network evolution patterns considering voltage overruns and harmonic exceedances according to claim 3, characterized in that, The process of constructing a single-resource grid connection scenario based on the penetration rate combination vector of various flexible resources in the distribution network and the node grid connection probability of various flexible resources at different nodes includes: Random selection Each node is used as the initial node in a single resource grid connection scenario, and the set probability of the initial node is calculated. Randomly select several nodes, change the access status of the resources corresponding to the selected nodes, and obtain the probability of the node set of new candidate scenarios; Based on the initial set probability and the new candidate scene set probability, the Markov chain state transition method is used to calculate the acceptance probability of the newly generated candidate scene and generate a random number. Based on the obtained acceptance probability and the generated random number, determine whether to accept the newly generated candidate scene; Based on the judgment results, a single resource grid connection scenario is obtained.

5. The method for classifying distribution network evolution patterns considering voltage overruns and harmonic exceedances according to claim 4, characterized in that, The process of determining whether to accept a newly generated candidate scene based on the obtained acceptance probability and the generated random number includes: If the random number is less than the acceptance probability, the state transitions, the candidate scenario is accepted as the new grid connection scenario, and the new grid connection scenario is used as the starting point for the next iteration; otherwise, the current scenario is retained.

6. The method for classifying distribution network evolution patterns considering voltage overruns and harmonic exceedances according to claim 1, characterized in that, The construction of the distribution network voltage over-limit matrix includes: Calculate the voltage deviation amplitude based on the voltage amplitude of each node at different time sections and the nominal voltage of the distribution network system; The voltage over-limit judgment coefficient is obtained based on the voltage deviation magnitude and the allowable voltage deviation of the distribution network; Based on the voltage deviation amplitude and the voltage over-limit judgment coefficient, a voltage over-limit matrix reflecting the distribution network under different time sections is constructed.

7. The method for classifying distribution network evolution patterns considering voltage overruns and harmonic exceedances according to claim 1, characterized in that, The construction of the distribution network harmonic exceedance matrix includes: Calculate the degree of voltage total harmonic distortion (THD) exceeding the standard and the judgment coefficient of the degree of voltage THD exceeding the standard, and construct the system voltage THD characteristic matrix; Obtain the voltage content of each harmonic of the system and construct the voltage content matrix of each harmonic of the system; Based on the system voltage total harmonic distortion rate characteristic matrix and the system harmonic voltage content matrix, a matrix reflecting the harmonic exceedance of the distribution network under different time sections is constructed.

8. The method for classifying distribution network evolution patterns considering voltage overruns and harmonic exceedances according to claim 1, characterized in that, The method of partitioning multi-resource grid-connected scenarios using the density clustering DBSCAN algorithm based on the distribution network voltage over-limit matrix includes: Based on the Hausdorff distance method, the similarity of voltage over-limit features of the scene is calculated; Set the neighborhood radius of the scene voltage similarity and the sample radius of the distribution network voltage over-limit matrix of the scene as a threshold of the number of voltage over-limit matrix samples of other scenes in the neighborhood of the scene voltage similarity; Randomly select a scene and calculate all scenes whose voltage over-limit feature similarity to the selected scene is less than or equal to the scene voltage similarity neighborhood radius; If the total number of all obtained scenarios is less than the threshold, the selected scenario is discarded, a new scenario is selected, and all scenarios with voltage limit violation feature similarity to the selected scenario less than or equal to the voltage similarity neighborhood radius of the scenario are calculated. The process continues to determine whether the total number of all obtained scenarios is less than the threshold. Otherwise, the selected scenario is taken as the core scenario, and other scenarios are determined based on the core scenario. If it is a core scenario, the core scenarios are grouped into a cluster; otherwise, it is determined whether the unselected scenarios are core scenarios, and the core scenarios are grouped into a cluster, until all scenarios are divided into different clusters according to the distribution network voltage over-limit matrix, thus completing one division of the distribution network evolution form.

9. The method for classifying distribution network evolution patterns considering voltage overruns and harmonic exceedances according to claim 8, characterized in that, Based on the initial partitioning result, the DBSCAN algorithm is used to perform a secondary partitioning of the multi-resource grid-connected scenario according to the distribution network harmonic exceedance matrix, including: In each cluster of the initial division, the harmonic similarity of the scenarios is calculated, and the DBSCAN algorithm is used to re-cluster all multi-resource grid-connected scenarios to achieve a secondary division of the distribution network evolution pattern.

10. A distribution network evolution morphology classification device considering voltage over-limit and harmonic exceedance, characterized in that, include: The grid connection scenario generation module is used to generate multi-resource grid connection scenarios for various flexible resources in the distribution network; The calculation module is used to perform power flow calculation and harmonic power flow calculation in multi-resource grid-connected scenarios, and to obtain the voltage amplitude, voltage harmonic content and total harmonic distortion rate of each node at different time sections. The power quality exceedance matrix construction module is used to construct the distribution network voltage exceedance matrix and the distribution network harmonic exceedance matrix based on the voltage amplitude, voltage harmonic content and total harmonic distortion rate of each node at different time sections. The grid connection scenario segmentation module is used to perform two-stage segmentation of multi-resource grid connection scenarios based on the distribution network voltage over-limit matrix and the distribution network harmonic over-limit matrix.

Images