A method combining Dijkstra algorithm and AFK-MC 2 Microgrid cluster division method based on clustering algorithm

Abstract

The present invention discloses a method combining Dijkstra algorithm and AFK-MC 2 A microgrid clustering method based on a clustering algorithm relates to the technical field of microgrid clustering, including deploying a collection network, continuously collecting and processing operation data in a microgrid according to the collection network; obtaining the shortest path between microgrid nodes by a Dijkstra algorithm, and constructing a feature vector in combination with the operation data; and using AFK‑MC to obtain the shortest path between microgrid nodes according to the feature vector. 2 The clustering algorithm generates preliminary microgrid cluster division results; the Dijkstra algorithm is used to re-optimize the preliminary division results to achieve dynamic division of microgrid clusters. The present invention deploys a comprehensive acquisition network and a real-time data processing system, combines the Dijkstra shortest path algorithm and AFK-MC 2 Clustering algorithm realizes efficient dynamic partitioning of microgrid clusters.

Description

Technical Field

[0001] The present invention relates to the technical field of microgrid cluster partitioning, in particular to a method combining Dijkstra algorithm and AFK-MC 2 Microgrid cluster division method based on clustering algorithm. Background Art

[0002] With the transformation of energy structure and the widespread application of renewable energy, microgrids are increasingly valued as an important way to achieve efficient energy utilization and optimize power system operation. Collaborative optimization, energy management and scheduling strategies among microgrid clusters are current research hotspots in the field of power systems. As an important part of smart grids, efficient clustering of microgrids is of great significance for optimizing power resource allocation, improving energy utilization efficiency and ensuring stable operation of power grids. Microgrid clustering involves grouping geographically dispersed microgrids according to certain characteristics or parameters for easy management and optimized operation.

[0003] Existing technologies are all based on static division, and most of them are one-time microgrid cluster division. Most scholars first use Ohm's law, equivalent impedance and other methods to measure the distance between microgrids, and then divide the microgrid clusters through clustering algorithms combined with the distance calculated in the previous step. It cannot support the scenario of adding or deleting a microgrid in the microgrid cluster. Most scholars use fixed distances and fixed numbers of microgrids to realize the division of microgrid clusters, and do not consider the need to add or delete microgrids in the future. Summary of the invention

[0004] In view of the problems existing in the existing microgrid cluster division, the present invention is proposed.

[0005] Therefore, the problem to be solved by the present invention is that it cannot support the scenario of adding or deleting a microgrid in a microgrid cluster. Most scholars implement the division of microgrid clusters based on a fixed distance and a fixed number of microgrids, and do not consider the need to add or delete microgrids in the future.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In the first aspect, the embodiment of the present invention provides a method combining Dijkstra algorithm and AFK-MC 2 A microgrid clustering method based on clustering algorithm, which includes:

[0008] Deploy a collection network, and continuously collect and process operation data in the microgrid according to the collection network;

[0009] Obtain the shortest path between microgrid nodes by using Dijkstra algorithm, and construct a feature vector in combination with the operation data;

[0010] According to the feature vector, AFK-MC 2 The clustering algorithm generates preliminary microgrid cluster division results;

[0011] The Dijkstra algorithm is used to re-optimize the preliminary division results and realize the dynamic division of microgrid clusters.

[0012] As the present invention combines Dijkstra algorithm and AFK-MC 2 A preferred solution of a microgrid cluster division method based on a clustering algorithm, wherein: the process of the Dijkstra algorithm includes:

[0013] Create a priority queue and set the distance estimate of the source node to 0;

[0014] Add the source node to a priority queue and initialize the distance estimates of all other nodes to positive infinity;

[0015] Set the predecessor node array;

[0016] When the priority queue is not empty, extract the node u with the smallest estimated distance from the priority queue, and for each adjacent node v of node u, calculate the path distance from the source node through u to v. If this path distance is less than the currently recorded d[v] value, update d[v] and the predecessor node π[v], and add v and its new estimated distance value to the priority queue;

[0017] When the priority queue is empty, the shortest path distances of all nodes have been determined, and a distance matrix is ​​obtained; at this time, the distance array d records the shortest path distance from the source node to each node, and the predecessor node array π is used to reconstruct these paths.

[0018] As the present invention combines Dijkstra algorithm and AFK-MC 2 A preferred solution for the microgrid clustering method based on clustering algorithm, wherein: the AFK-MC 2 The clustering algorithm generates preliminary microgrid cluster division results including:

[0019] Perform feature extraction based on the distance matrix to obtain a comprehensive feature vector;

[0020] Performing adaptive parameter initialization on the comprehensive feature vector;

[0021] Calculate the membership of each node to the cluster center according to the initialized feature vector;

[0022] Adjust the connectivity between nodes based on physical connections and energy exchange;

[0023] The clustering quality is evaluated, and parameters are adjusted according to the clustering quality, and the effect is evaluated.

[0024] As the present invention combines Dijkstra algorithm and AFK-MC 2 A preferred solution of the microgrid cluster division method of the clustering algorithm, wherein: the calculation of the membership of each node to the cluster center is shown as follows:

[0025]

[0026] Where m is the fuzzification parameter, ‖·‖ represents the vector norm, t is the number of iterations, i is the node, c is the cluster center, C is the number of clusters in the fuzzy C-means algorithm, and u is the t i,c Indicates the membership of node i to cluster center c at the tth iteration, ||X i -μ c || represents the distance between node i and cluster center c, that is, the feature vector X of node i i and the eigenvector μ of the cluster center c c The distance between ||X i -μ c′ || represents the distance between node i and cluster center c′, that is, the feature vector X of node i i and the eigenvector μ of the cluster center c′ c′ The distance between them, c′ represents the temporary cluster center index, which is used for cumulative calculation.

[0027] As the present invention combines Dijkstra algorithm and AFK-MC 2 A preferred solution of the microgrid cluster division method of the clustering algorithm, wherein: the connectivity between the nodes is adjusted as shown in the following formula:

[0028] k ij =w physical ·d ij -α +w energy ·e β ij

[0029] Among them, d ij is the Dijkstra distance, e ij is the energy exchange, w physical ,w energy is the weight factor, a and β are adjustment parameters.

[0030] As the present invention combines Dijkstra algorithm and AFK-MC2 A preferred solution of a microgrid cluster partitioning method based on a clustering algorithm, wherein: the use of the Dijkstra algorithm to re-optimize the preliminary partitioning results includes

[0031] Take each node in the cluster as the starting point and calculate the shortest path to all other nodes using the Dijkstra algorithm;

[0032] Based on the calculation results, a distance matrix is ​​constructed to reflect the electrical distance between nodes within the cluster;

[0033] Analyzing the distance matrix to identify edge nodes that are weakly associated with the current cluster;

[0034] For the edge node, calculating the distance between the edge node and an adjacent cluster node, and evaluating whether the edge node should be reallocated to an adjacent cluster;

[0035] If a better allocation scheme is found, the cluster boundaries are adjusted and the edge nodes are reallocated.

[0036] As the present invention combines Dijkstra algorithm and AFK-MC 2 A preferred solution of a microgrid cluster division method using a clustering algorithm, wherein: the Dijkstra algorithm also includes a database for storing the operating data; the database adopts a graph structure storage mode;

[0037] The graph structure storage mode includes:

[0038] Map each physical node in the microgrid to a vertex in the graph. Each vertex has a unique identifier and attributes, and the physical connections between nodes are mapped to edges in the graph, each edge carrying connection information.

[0039] The mapping is a dynamic mapping. When a new node or line failure occurs, it is immediately reflected in the graph structure. The instant reflection is specifically to continuously monitor the network status through the intelligent sensor network distributed in the microgrid. When a new node is connected or a line failure is detected, these sensors will immediately generate an event report and send it to the central control system through the communication network. For the new node, a new graph vertex is automatically generated, a unique identifier is assigned to it, and related attributes are initialized. At the same time, the connection relationship between the new node and the existing network is analyzed, the corresponding edge is created, and the initial weight is calculated. For a line failure, the system will immediately mark the corresponding edge as unavailable and trigger a dynamic adjustment strategy.

[0040] The dynamic adjustment strategy specifically includes quickly adjusting the routing strategy of the affected area without recalculating the entire network, and improving the graph structure according to the routing strategy.

[0041] In a second aspect, the embodiment of the present invention provides a method combining Dijkstra algorithm and AFK-MC 2 A microgrid cluster partitioning system based on clustering algorithm, comprising:

[0042] A data acquisition and processing module is used to deploy an acquisition network, and continuously acquire and process the operation data in the microgrid according to the acquisition network;

[0043] The dynamic clustering module is used to obtain the shortest path between microgrid nodes through the Dijkstra algorithm and construct a feature vector based on the operation data; according to the feature vector, AFK-MC is used 2 The clustering algorithm generates preliminary microgrid cluster division results;

[0044] The optimization module is used to re-optimize the preliminary partitioning results using the Dijkstra algorithm to achieve dynamic partitioning of the microgrid cluster.

[0045] In a third aspect, an embodiment of the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, wherein: when the processor executes the computer program, the above-mentioned combination of Dijkstra algorithm and AFK-MC is implemented. 2 Any step of the microgrid clustering method based on clustering algorithm.

[0046] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, the above-mentioned combination of Dijkstra algorithm and AFK-MC is implemented. 2 Any step of the microgrid clustering method based on clustering algorithm.

[0047] The beneficial effect of the present invention is that by deploying a comprehensive collection network and a real-time data processing system, the Dijkstra shortest path algorithm and AFK-MC 2 The clustering algorithm realizes the efficient dynamic division of microgrid clusters. This method can not only collect and analyze multi-dimensional operation data in real time, but also adaptively adjust the cluster structure according to network topology changes and operation conditions. By optimizing data storage and processing strategies and combining graph structure storage mode, the scheme significantly improves the efficiency of algorithm execution. Finally, through the iterative optimization process, the precise division of microgrid clusters is achieved, the overall reliability, flexibility and economy of the system are improved, and strong technical support is provided for the intelligent management and optimized operation of microgrids. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0049] Figure 1 To combine Dijkstra algorithm with AFK-MC 2 Flowchart of the overall scheme of the microgrid cluster division method based on clustering algorithm.

[0050] Figure 2 To combine Dijkstra algorithm with AFK-MC 2 AFK-MC based on clustering algorithm for adaptive mechanism of microgrid cluster division method 2 Cluster partitioning diagram.

[0051] Figure 3 To combine Dijkstra algorithm with AFK-MC 2 Flowchart of Dijkstra algorithm for microgrid cluster partitioning method based on clustering algorithm. DETAILED DESCRIPTION

[0052] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0053] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0054] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0055] The present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0056] At the same time, in the description of the present invention, it should be noted that the directions or positional relationships indicated by the terms "upper, lower, inner and outer" are based on the directions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as limiting the present invention. In addition, the terms "first, second or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0057] In the present invention, unless otherwise clearly specified and limited, the terms "install, connect, connect" should be understood in a broad sense, for example: it can be a fixed connection, a detachable connection or an integral connection; it can also be a mechanical connection, an electrical connection or a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0058] Example 1

[0059] Reference Figure 1 to Figure 3 , which is the first embodiment of the present invention, and provides a method combining Dijkstra algorithm and AFK-MC 2 The microgrid cluster division method based on clustering algorithm includes:

[0060] It should be noted that the Dijkstra algorithm is the Dijkstra algorithm, and AFK-MC 2 The clustering algorithm is AFK-MC 2 Clustering algorithms.

[0061] S1: deploying a data collection network, and continuously collecting and processing the operating data in the microgrid according to the data collection network.

[0062] Specifically, first, a comprehensive network of collectors is deployed in the microgrid. These collectors are not limited to traditional current, voltage and power data collection, but also include a wider range of parameters such as communication delay between nodes, link quality indicators, the status of energy storage devices, renewable energy generation forecasts, etc.

[0063] These collectors adopt a distributed architecture and achieve real-time data transmission through low-power wide area network (LPWAN) technology to ensure the timeliness and integrity of the data. Next, an adaptive algorithm based on artificial intelligence is used in the data preprocessing stage to perform data cleaning and anomaly detection. It can not only identify and process common data anomalies, but also dynamically adjust the cleaning strategy through machine learning models to adapt to the characteristics of different microgrids. In the process of data standardization, a new multi-scale standardization technology is introduced. This technology takes into account the data characteristics of different time scales (such as seconds, minutes, and hours), which provides a basis for the subsequent Dijkstra algorithm and AFK-MC 2 The clustering algorithm provides a richer information foundation. This multi-level and multi-angle data collection and preprocessing method provides a basis for subsequent Dijkstra shortest path calculation and AFK-MC 2 Clustering provides high-quality, multi-dimensional input data, which significantly improves the accuracy and adaptability of microgrid cluster division.

[0064] Furthermore, for the data collected and processed in real time, the time series database technology is used to optimize the large-scale and high-frequency data flow in the microgrid environment. This real-time database is not just a simple data storage facility, but an intelligent data management platform. It can receive and store multi-dimensional real-time data that has been pre-processed by the upper layer, including power parameters, network topology information, equipment status, etc.

[0065] In order to support the efficient execution of the Dijkstra algorithm, the database adopts a graph structure storage mode, which directly maps the node relationships in the microgrid into a graph data structure, greatly accelerating the speed of path calculation.

[0066] Preferably, the graph structure storage mode first maps each physical node in the microgrid (such as power stations, substations, load centers, etc.) to a vertex in the graph. Each vertex has a unique identifier and a set of attributes. The physical connections between nodes are mapped to edges in the graph. Each edge carries connection information (line impedance, maximum transmission capacity, current load).

[0067] Specifically, the mapping is not just static, but dynamically updated. The system monitors the topological changes of the microgrid in real time. When a new node is added or a line fails, it is immediately reflected in the graph structure. Specifically, the immediate reflection is through the distributed intelligent sensor network in the microgrid to continuously monitor the network status (access signals of new devices and operating parameters of existing devices). When a new node is detected or a line failure is detected, these sensors will immediately generate an event report and send it to the central control system through the communication network. For the new node, a new graph vertex is automatically generated, a unique identifier is assigned to it, and related attributes are initialized. At the same time, the connection relationship between the new node and the existing network is analyzed, the corresponding edge is created, and the initial weight is calculated. For a line failure, the system will immediately mark the corresponding edge as unavailable and trigger a dynamic routing recalculation process. Without recalculating the entire network, the routing strategy of the affected area can be quickly adjusted, and the graph structure can be improved according to the routing strategy.

[0068] S2: Obtain the shortest path between microgrid nodes through Dijkstra algorithm, and construct a feature vector based on the operation data.

[0069] Specifically, the Dijkstra algorithm is used to calculate the shortest transmission path between all nodes in the power grid. This step not only takes into account the physical connection, but also factors such as transmission loss. Then, according to the above steps, the multi-dimensional parameter data of the power grid operation is collected and integrated, including but not limited to the power generation power, load demand, voltage level, power factor, etc. of each node. Combining these operating parameters with the shortest path information, we can construct a high-dimensional feature vector that comprehensively describes the position and operating status of each node in the network.

[0070] Specifically, the process of Dijkstra algorithm includes initialization, process selection and update, and termination.

[0071] initialization:

[0072] Create a priority queue to store nodes and their current distance estimates, set the source node's distance estimate to 0, and add it to the priority queue. Initialize all other nodes' distance estimates to positive infinity (indicating that no path has been found yet). Set the predecessor node array to record the previous node on the shortest path.

[0073] When the priority queue is not empty, perform the following steps:

[0074] Extract the node u with the smallest estimated distance from the priority queue. This is the node closest to the source node among the currently unvisited nodes. For each adjacent node v of node u, calculate the path distance from the source node through u to v, that is, d[u]+weight(u,v). If this path distance is less than the currently recorded d[v] value (that is, a shorter path is found), update d[v] and the predecessor node π[v], and add v and its new estimated distance value to the priority queue.

[0075] When the priority queue is empty, the shortest path distances of all nodes have been determined. At this point, the distance array d records the shortest path distances from the source node to each node, and the predecessor node array π can be used to reconstruct these paths. Finally, the distance matrix is ​​obtained.

[0076] Compared with the Dijkstra algorithm without priority queue, the use of priority queue can significantly reduce the time complexity of selecting the minimum distance node each time from O(V) to O(logV). In addition, in large-scale graphs with a large number of nodes, the use of priority queue makes the Dijkstra algorithm more efficient. Combined with dynamic data and real-time calculation, the priority queue can flexibly adapt to changes in edge weights or structures in the graph, thereby improving the practical application value of the algorithm.

[0077] S3: According to the feature vector, use AFK-MC 2 The clustering algorithm generates preliminary microgrid cluster division results.

[0078] AFK-MC 2 The clustering algorithm engine is the core component of the microgrid clustering method. It uses the comprehensive analysis of real-time data and historical data to dynamically optimize the clustering structure of the microgrid. Specifically, the engine first collects real-time grid operation data from the acquisition network and historical operation data stored in the database. It analyzes these multi-dimensional and multi-time scale data to identify dynamic change patterns and potential clustering structures in the grid. It can adaptively adjust clustering parameters, such as the number of clusters and the location of cluster centers, to adapt to real-time changes in grid status.

[0079] Through this flexible clustering strategy, the system can achieve optimal configuration of nodes and power resources. For example, a larger microgrid group may be formed in certain periods to improve the overall efficiency of the system, while smaller microgrid units may be divided in other periods to enhance local self-sufficiency. This dynamic adjustment can also help the system better respond to emergencies, such as equipment failure or supply and demand imbalances under extreme weather conditions.

[0080] Optimized, adaptive AFK-MC 2 The clustering steps are as follows:

[0081] For feature extraction, assume that the power grid has N nodes, and the features of each node include but are not limited to Pi (node ​​power), ΔLi (load change rate), etc. The distance matrix obtained based on the Dijkstra algorithm is D∈RN×N. The comprehensive feature vector is:

[0082] X=[P 1 ,ΔL 1 ,...,P N ,ΔL N ,D 1,1 ...,D N,N ] T ∈R N×(2+N)

[0083] The distance matrix D∈R^(N×N) obtained by introducing the Dijkstra algorithm not only takes into account direct features such as node power Pi and load change rate ΔPi, but also takes into account the topological relationship between nodes.

[0084] Furthermore, adaptive parameter initialization is performed. The adaptive parameter initialization step introduces the total load level Ptotal and the proportion of renewable energy Prenew as influencing factors. The initial cluster number K and the membership threshold ε are determined by the function f. These parameters provide initial conditions for the iterative process:

[0085] C=f(L total ,R renew ),θ=g(L total ,R renew )

[0086] Furthermore, the membership of each node to each cluster center is calculated based on the initialized vector, and adaptive weights are introduced. Specifically,

[0087] At the tth iteration, the cluster center is μc∈R(2+N), and the membership degree μ of each node i to the cluster center c is i ,c t , calculated using the fuzzy C-means (FCM) formula, while incorporating the adaptive weight w i (Adjusted based on connectivity):

[0088]

[0089] Where m is the fuzzification parameter, ‖·‖ represents the vector norm, t is the number of iterations, i is the node, c is the cluster center, C is the number of clusters in the fuzzy C-means algorithm, and u is the t i,c Indicates the membership of node i to cluster center c at the tth iteration, ||X i -μ c|| represents the distance between node i and cluster center c, that is, the feature vector X of node i i and the eigenvector μ of the cluster center c c The distance between ||X i -μ c′ || represents the distance between node i and cluster center c′, that is, the feature vector X of node i i and the eigenvector μ of the cluster center c′ c′ The distance between them, c′ represents the temporary cluster center index, which is used for cumulative calculation.

[0090] Adjust the connectivity between nodes according to the physical connection and energy exchange, thereby affecting the adaptive weight in the membership calculation:

[0091] k ij =w physical ·d ij -α +w energy ·e β ij

[0092] Among them, d ij is the Dijkstra distance, e ij is the energy exchange, w physical ,w energy is the weight factor, a and β are adjustment parameters.

[0093] After each iteration, the clustering quality is evaluated and the parameters are adjusted accordingly. If the clustering quality is not optimal or the number of iterations is not full, it triggers parameter adjustment and starts a new round of iterations.

[0094]

[0095] If S is not optimal or the number of iterations is not full, adjust μ, c, θ, etc. and enter the next round of iteration.

[0096] Finally, evaluate the quality of the final clustering result. It may affect whether you need to rerun the algorithm or adjust the initial parameters. As follows:

[0097] If s(i)>0, it means that object i prefers its cluster rather than the nearest other clusters; if it is close to 1, it means that the cluster division is very good; if it is close to 0, it means that object i is at the boundary and it is difficult to determine its belonging; if s(i)<0, it may indicate that the cluster division is poor and object i may be misclassified.

[0098] S4: Use the Dijkstra algorithm to reoptimize the preliminary partitioning results and realize the dynamic partitioning of microgrid clusters.

[0099] When the division results of the microgrid cluster change, resulting in a change in the cluster structure, the system will trigger the real-time response mechanism of the Dijkstra algorithm. The core of this mechanism is to quickly adapt to the new network topology to ensure the continuous optimization of the energy transmission path. Then the clustering results are continuously improved through adaptive control logic.

[0100] Specifically, the Dijkstra algorithm is used to analyze and reconstruct the 2 Initial microgrid clusters formed by clustering. The shortest path characteristics of the Dijkstra algorithm are used to re-evaluate and adjust cluster boundaries, including:

[0101] Take each node in the cluster as the starting point and run the Dijkstra algorithm to calculate the shortest path to all other nodes;

[0102] Based on the calculation results, a distance matrix is ​​constructed to reflect the electrical distance between nodes within the cluster;

[0103] Analyze the distance matrix to identify edge nodes that are weakly associated with the current cluster (i.e., farther away);

[0104] For these edge nodes, their distances to adjacent cluster nodes are calculated to evaluate whether they should be reallocated to adjacent clusters.

[0105] If a better allocation scheme is found, the cluster boundaries are adjusted and the edge nodes are reallocated;

[0106] Repeat the above steps until the boundaries of all clusters become stable or the preset number of iterations is reached.

[0107] This combination method retains the AFK-MC 2 The algorithm's advantages in processing large-scale data are combined with the strengths of the Dijkstra algorithm in path optimization, thereby achieving flexible, efficient and dynamic division of microgrid clusters and improving the reliability and economy of the entire system.

[0108] Furthermore, this embodiment also provides a method combining Dijkstra algorithm and AFK-MC 2 The microgrid cluster division system based on clustering algorithm includes:

[0109] A data acquisition and processing module is used to deploy an acquisition network, and continuously acquire and process the operation data in the microgrid according to the acquisition network;

[0110] The dynamic clustering module is used to obtain the shortest path between microgrid nodes through the Dijkstra algorithm and construct a feature vector based on the operation data; according to the feature vector, AFK-MC is used 2 The clustering algorithm generates preliminary microgrid cluster division results;

[0111] The optimization module is used to re-optimize the preliminary partitioning results using the Dijkstra algorithm to achieve dynamic partitioning of the microgrid cluster.

[0112] This embodiment also provides a computer device suitable for combining the Dijkstra algorithm with the AFK-MC 2 The method for dividing microgrid clusters based on clustering algorithm comprises a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the combination of Dijkstra algorithm and AFK-MC proposed in the above embodiment. 2 Microgrid cluster division method based on clustering algorithm.

[0113] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0114] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the above embodiment combines the Dijkstra algorithm with the AFK-MC algorithm. 2 Microgrid cluster division method based on clustering algorithm.

[0115] The storage medium proposed in this embodiment and the data storage method proposed in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be found in the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0116] Example 2

[0117] This is the second embodiment of the present invention. This embodiment provides a microgrid cluster division method that combines the Dijkstra algorithm with the AFK-MC 2 clustering algorithm. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0118] In order to verify the effectiveness and advantages of the proposed dynamic partitioning method for microgrid clusters, this embodiment conducts detailed experiments in a complex microgrid system consisting of 30 nodes. The microgrid system includes multiple types of power generation units (such as photovoltaic power generation, wind power generation, and traditional thermal power generation), multiple load centers, and distributed energy storage devices.

[0119] First, according to step S1, a comprehensive collection network is deployed. High-precision smart sensors are installed on each node, which can not only collect traditional power parameters (such as voltage, current, active power and reactive power), but also measure the communication delay between nodes, link quality indicators, the state of charge (SOC) of energy storage devices, and the real-time power generation and short-term forecast values ​​of renewable energy. In order to ensure the real-time and reliability of data transmission, LoRaWAN technology is used to build a low-power wide area network, realizing efficient communication between the collector and the central control system.

[0120] The optimized Dijkstra algorithm is used to calculate the shortest paths between all nodes. The algorithm not only considers the physical distance, but also factors such as line impedance and current load level in the weight calculation, thereby obtaining an "electrical distance" that is more in line with the actual operating conditions. Based on this shortest path information, combined with multi-dimensional operating data collected in real time, a 31-dimensional feature vector (including node power, load change rate, voltage level, power factor, SOC, shortest path distance, etc.) is constructed.

[0121] Utilizing the Improved AFK-MC 2 The clustering algorithm performs preliminary microgrid clustering. The algorithm dynamically adjusts clustering parameters by analyzing real-time and historical data. For example, during peak hours, the algorithm tends to form larger microgrid groups to improve the overall efficiency of the system; while during off-peak hours, it tends to divide smaller microgrid units to enhance local self-sufficiency.

[0122] Finally, the optimized Dijkstra algorithm is called again to optimize the preliminary partitioning results. By analyzing the electrical distances of nodes within the cluster, edge nodes are identified and evaluated to see whether they should be reallocated to adjacent clusters. This process is repeated until the boundaries of all clusters become stable.

[0123] In order to quantitatively evaluate the performance of this method, tests were conducted under normal operating conditions and fault conditions, and compared with the traditional static partitioning method. The test results are shown in the following table:

[0124] Table 1 Experimental data table

[0125]

[0126]

[0127] By analyzing the above experimental data, we can draw the following important conclusions:

[0128] Convergence speed: In all test scenarios, the average convergence time of the method of the present invention is significantly lower than that of the traditional static method. Under normal operating conditions, the convergence time of the present method is 30.4% shorter than that of the traditional method, which is mainly due to the application of the optimized Dijkstra algorithm and the graph structure storage mode. Especially in the case of multi-node failure, the convergence time advantage of the present method is more obvious, which is 48.4% faster than the traditional method, reflecting the high efficiency of the present method in complex dynamic environments.

[0129] Partition accuracy: The method of the present invention maintains a high partition accuracy under various working conditions. Even in the most challenging multi-node failure situation, its accuracy still reaches 88.5%, which is 29.6 percentage points higher than the traditional method. This fully demonstrates the superiority of the method integrating multi-dimensional data and dynamic optimization strategy.

[0130] System stability: This method significantly improves the overall stability of the microgrid system. Under normal operating conditions, the stability index reaches 0.89, which is 17.1% higher than the traditional method. More notably, under extreme weather conditions, this method can still maintain a high stability of 0.83, while the traditional method drops to 0.53, which fully demonstrates the robustness of this method.

[0131] Energy efficiency: Thanks to precise cluster partitioning and dynamic optimization, this method achieves higher energy efficiency under various conditions. Under normal operating conditions, the efficiency is improved by 8.3 percentage points to 86.5%. Even in the case of multiple node failures, the energy efficiency of this method remains at 79.4%, 18.1 percentage points higher than the traditional method.

[0132] Load balancing: This method performs well in maintaining load balancing. In all test scenarios, the load balancing degree of this method is higher than that of traditional methods. In particular, in the case of a single node failure, the load balancing degree of this method is 0.79, which is 0.24 higher than that of the traditional method, indicating that this method can better cope with local failures and maintain the overall balance of the system.

[0133] Dynamic adaptability: The advantages of this method in dynamic adaptability are particularly prominent. In all test scenarios, the dynamic adaptability score of this method is above 4.3, while the highest score of the traditional method is only 2.1. This fully proves that this method can quickly respond to changes in network topology and fluctuations in operating status, and realize flexible adjustment of microgrid clusters.

[0134] In summary, the dynamic partitioning method of microgrid clusters proposed in the present invention is significantly superior to the traditional static partitioning method in multiple key indicators such as convergence speed, partitioning accuracy, system stability, energy efficiency, load balancing and dynamic adaptability. Especially in the face of complex dynamic environments (such as multi-node failures and extreme weather conditions), this method shows strong robustness and adaptability. These advantages mainly come from the fact that this method integrates multidimensional data acquisition, intelligent data processing, optimized path algorithm and adaptive clustering strategy. By perceiving the network status in real time and quickly adjusting the partitioning strategy, this method provides strong technical support for the intelligent management and optimized operation of microgrids, and is expected to play an important role in improving the reliability, economy and environmental friendliness of the power grid.

[0135] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method combining Dijkstra algorithm and AFK-MC 2 The microgrid cluster division method based on clustering algorithm is characterized by: include, Deploy a collection network, and continuously collect and process operation data in the microgrid according to the collection network; Obtain the shortest path between microgrid nodes by using Dijkstra algorithm, and construct a feature vector in combination with the operation data; According to the feature vector, AFK-MC 2 The clustering algorithm generates preliminary microgrid cluster division results; Use Dijkstra algorithm to re-optimize the preliminary partitioning results and realize dynamic partitioning of microgrid clusters; The Dijkstra algorithm also includes a database storing the operating data; The database adopts a graph structure storage mode; The graph structure storage mode includes: Map each physical node in the microgrid to a vertex in a graph structure. Each vertex has a unique identifier and attributes, and the physical connections between nodes are mapped to edges in the graph structure, with each edge carrying connection information. The mapping is a dynamic mapping. When a new node or line failure occurs, it is immediately reflected in the graph structure. The instant reflection is specifically to continuously monitor the network status through the intelligent sensor network distributed in the microgrid. When a new node is connected or a line failure is detected, these sensors will immediately generate an event report and send it to the central control system through the communication network. For the new node, a new graph vertex is automatically generated, a unique identifier is assigned to it, and related attributes are initialized. At the same time, the connection relationship between the new node and the existing network is analyzed, the corresponding edge is created, and the initial weight is calculated. For a line failure, the system will immediately mark the corresponding edge as unavailable and trigger a dynamic adjustment strategy. The dynamic adjustment strategy specifically includes quickly adjusting the routing strategy of the affected area without recalculating the entire network, and improving the graph structure according to the routing strategy.

2. The method of combining Dijkstra algorithm and AFK-MC as claimed in claim 1 2 The microgrid cluster division method based on clustering algorithm is characterized by: The process of the Dijkstra algorithm includes: Create a priority queue and set the distance estimate of the source node to 0; Add the source node to a priority queue and initialize the distance estimates of all other nodes to positive infinity; Set the predecessor node array; When the priority queue is not empty, extract the node u with the smallest estimated distance from the priority queue, and for each adjacent node v of node u, calculate the path distance from the source node through u to v. If this path distance is less than the currently recorded d[v] value, update d[v] and the predecessor node π[v], and add v and its new estimated distance value to the priority queue; When the priority queue is empty, the shortest path distances of all nodes have been determined, and a distance matrix is ​​obtained; at this time, the distance array d records the shortest path distance from the source node to each node, and the predecessor node array π is used to reconstruct these paths.

3. The method of combining Dijkstra algorithm and AFK-MC as claimed in claim 2 2 The microgrid cluster division method based on clustering algorithm is characterized by: The AFK-MC 2 The clustering algorithm generates preliminary microgrid cluster division results including: Perform feature extraction based on the distance matrix to obtain a comprehensive feature vector; Performing adaptive parameter initialization on the comprehensive feature vector; Calculate the membership of each node to the cluster center according to the initialized feature vector; Adjust the connectivity between nodes based on physical connections and energy exchange; The clustering quality is evaluated, and parameters are adjusted according to the clustering quality, and the effect is evaluated.

4. The method of combining Dijkstra algorithm and AFK-MC as claimed in claim 3 2 The microgrid cluster division method based on clustering algorithm is characterized by: The calculation of the membership of each node to the cluster center is shown in the following formula: Where m is the fuzzification parameter, ‖·‖ represents the vector norm, t is the number of iterations, i is the node, c is the cluster center, C is the number of clusters in the fuzzy C-means algorithm, and u is the t i,c Indicates the membership of node i to cluster center c at the tth iteration, ||X i -μ c || represents the distance between node i and cluster center c, that is, the feature vector X of node i i and the eigenvector μ of the cluster center c c The distance between ||X i -μ c′ || represents the distance between node i and cluster center c′, that is, the feature vector X of node i i and the eigenvector μ of the cluster center c′ c′ The distance between them, c′ represents the temporary cluster center index, which is used for cumulative calculation.

5. The method of combining Dijkstra algorithm and AFK-MC as claimed in claim 3 2 The microgrid cluster division method based on clustering algorithm is characterized by: The connection degree between the nodes is adjusted as shown in the following formula: k ij =w physical ·d ij -α +w energy ·e β ij Among them, d ij is the Dijkstra distance, e ij is the energy exchange, w physical ,w energy is the weight factor, a and β are adjustment parameters, i represents the i node, and j represents the j node.

6. The method of combining Dijkstra algorithm and AFK-MC as claimed in claim 5 2 The microgrid cluster division method based on clustering algorithm is characterized by: The use of the Dijkstra algorithm to reoptimize the preliminary partitioning results includes: Take each node in the cluster as the starting point and calculate the shortest path to all other nodes using the Dijkstra algorithm; Based on the calculation results, a distance matrix is ​​constructed to reflect the electrical distance between nodes within the cluster; Analyzing the distance matrix to identify edge nodes that are weakly associated with the current cluster; For the edge node, calculating the distance between the edge node and an adjacent cluster node, and evaluating whether the edge node should be reallocated to an adjacent cluster; If a better allocation scheme is found, the cluster boundaries are adjusted and the edge nodes are reallocated.

7. A method combining Dijkstra algorithm and AFK-MC 2 A microgrid cluster partitioning system based on a clustering algorithm, based on any one of claims 1 to 6, combining the Dijkstra algorithm and AFK-MC 2 The microgrid cluster division method based on clustering algorithm is characterized by: include, A data acquisition and processing module, used to deploy an acquisition network, continuously acquire and process the operating data in the microgrid according to the acquisition network; The dynamic clustering module is used to obtain the shortest path between microgrid nodes through the Dijkstra algorithm and construct a feature vector based on the operation data; according to the feature vector, AFK-MC is used 2 The clustering algorithm generates preliminary microgrid cluster division results; The optimization module is used to re-optimize the preliminary partitioning results using the Dijkstra algorithm to achieve dynamic partitioning of the microgrid cluster.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the combination of Dijkstra algorithm and AFK-MC according to any one of claims 1 to 6 is implemented. 2 Steps of the microgrid clustering method based on clustering algorithm.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the combination of Dijkstra algorithm and AFK-MC according to any one of claims 1 to 6 is realized. 2 Steps of the microgrid clustering method based on clustering algorithm.

Images