Intelligent emergency dispatching optimization management method for wind power station based on heuristic algorithm

Abstract

The invention discloses a heuristic algorithm-based intelligent emergency scheduling optimization management method for a wind power plant station, and belongs to the technical field of emergency management of the wind power plant station. The problems that most of existing wind power stations depend on traditional manual management and paper plans, the updating period of the plans is long, decision makers are difficult to obtain accurate information due to the fact that all emergency resources are lack of an effective information sharing mechanism and real-time data support, and a system is difficult to provide visual emergency event progress display are solved. According to the technical scheme provided by the invention, the whole framework adopts main station and substation cooperative computing, information of a wind power plant station is converted into digital characteristics, and a swarm intelligent optimization algorithm-genetic algorithm is utilized, so that the intelligent level of emergency management of the wind power plant station is improved, and effective scheduling and optimal management of emergency resources are realized; the emergency response speed and the processing efficiency are improved, meanwhile, a real-time monitoring interface is provided, and event tracking and visual display of the processing process are supported.

Description

Technical Field

[0001] The present invention relates to the technical field of wind farm emergency management, and in particular to a wind farm intelligent emergency dispatch optimization management method based on a heuristic algorithm. Background Art

[0002] With the rapid development and large-scale application of wind power technology, wind farms, as an important clean energy production facility, are facing increasingly complex management and operation challenges. Especially in the field of emergency management, the emergency response capability of wind farms directly affects the stability and economic benefits of the overall power grid. However, the emergency management of many wind farms still relies on traditional manual operations and paper plans, which is particularly backward in the context of modern informatization and cannot effectively respond to complex emergencies brought about by the scale of wind farms and the dynamically changing external environment.

[0003] First, traditional emergency management relies on paper plans. Most of these plans are static documents, lacking flexibility and the ability to be dynamically updated. In actual operations, the update cycle of the plan is long, and it is difficult to quickly respond to changes in emergencies. Since the content of the plan usually needs to go through a complicated approval process, many wind farm plans cannot reflect the latest risk assessment and resource allocation in emergency situations. This lagging emergency management method appears passive in actual emergencies, often missing the best time to deal with it, resulting in an inability to respond to emergencies quickly and effectively.

[0004] Secondly, under the traditional management system, coordination between multiple management levels and functional departments is difficult, and the information island phenomenon is serious. Each department often acts independently and lacks a unified information sharing and dispatching platform, resulting in inefficient emergency resource allocation. In wind farm emergency management, emergency resources include technicians, maintenance equipment, and communication equipment. Since the information of various departments cannot be effectively synchronized, the allocation of resources is often delayed, which in turn affects the speed and efficiency of the overall emergency response. This management method that lacks information integration and sharing makes it difficult to improve the overall efficiency of emergency management.

[0005] Furthermore, the current emergency management system generally lacks the integration and analysis capabilities of real-time data. Traditional management methods rely more on experience and past cases rather than analysis based on real-time data. This makes it difficult for decision makers to quickly obtain accurate on-site information and adjust emergency strategies in a timely manner. Especially in wind farm emergency management, the on-site environment is complex and changeable. The lack of real-time data often means that emergency decisions cannot be effectively adjusted based on actual conditions, which may lead to emergency plans not meeting actual conditions and even exacerbating the complexity and severity of the problem.

[0006] In addition, the visualization capabilities of traditional emergency management systems are insufficient. Emergency management involves the collaboration of multiple links and departments. In actual emergency operations, managers need to grasp the progress of events and the dispatch of emergency resources in real time. However, the current system is difficult to provide an intuitive and comprehensive display of the progress of emergency events. Most emergency management systems can only present basic text information or tabular data, lacking a visual display of the overall emergency situation. This low-visualization management method makes it difficult for managers to intuitively perceive the overall emergency progress and situation, which in turn affects the accuracy and timeliness of emergency decision-making.

[0007] In response to the above problems, although some traditional technical solutions have attempted to optimize wind farm emergency management, the results are limited. For example, some wind farms have introduced emergency plan optimization methods based on historical data. This method can adjust the content of the plan according to past emergency events. However, this method is still based on static historical data and lacks the ability to dynamically adjust real-time information. In addition, some wind farms have introduced simplified monitoring systems for real-time monitoring of the operating status of wind turbines. However, these systems are mostly centered on equipment status and cannot provide a global perspective on overall emergency management, nor can they support cross-departmental coordination and scheduling.

[0008] Some plans also attempt to improve emergency response capabilities by increasing emergency training and drills. Although these measures can indeed improve the emergency response capabilities of personnel, they still fail to fundamentally solve the problems of information islands and untimely resource dispatch. Because emergency events often occur suddenly, there is a large gap between the scenarios simulated by traditional drills and the actual situation, and they cannot fully meet the needs of wind farms in complex emergency situations.

[0009] In general, the main bottlenecks of wind farm emergency management are inflexible emergency plan management, insufficient information sharing, lack of real-time decision support, and low system visualization. Although some technical means have been used to try to solve some of the problems, most of them are aimed at optimizing a single aspect and have failed to form a systematic overall solution. In the face of complex emergency events, wind farm emergency management is still lagging and passive, and cannot achieve truly efficient emergency response.

[0010] Therefore, in order to better cope with the emergency management needs of wind farms in the context of large-scale application, there is an urgent need for an emergency dispatch optimization management method that integrates high automation, real-time data monitoring, intelligent fault diagnosis and decision support. This method should not only be able to dynamically update plan management and realize cross-departmental information integration and sharing, but also provide decision makers with fast and accurate decision support based on real-time data. At the same time, the method should also have a highly visual operation interface to facilitate managers to control the overall situation of emergency events in real time. This new emergency management method will effectively solve the shortcomings of traditional solutions, improve the level of intelligence in wind farm emergency management, and ensure the safe, stable and efficient operation of wind power resources. Summary of the invention

[0011] In view of the problems existing in the research of the prior art, the present invention proposes an intelligent emergency dispatch optimization management method for wind farms based on heuristic algorithms. By constructing a three-dimensional visualization model of the wind farm, including all facilities, equipment and the surrounding environment in the wind farm, a digital network is established, and the main and substation collaborative computing method is adopted to convert the information of the wind farm into digital features. The group intelligence optimization algorithm, namely the genetic algorithm, is used to improve the intelligence level of emergency management of wind farms.

[0012] To achieve the above object, the present invention is implemented through the following technical solutions:

[0013] The present invention proposes a wind farm intelligent emergency dispatch optimization management method based on a heuristic algorithm, comprising the following steps:

[0014] Step 1: Use image acquisition equipment to collect visual information, including topography, wind turbine towers, substations, cable lines, crowd numbers, road codes, road accessibility, and road risk levels; use temperature acquisition equipment to collect temperature information, including the temperature values ​​inside and outside the wind turbine station; use measuring instruments to collect equipment operating status, including voltage, current, and power;

[0015] Step 2: Based on the collected visual information, build a 3D model of important facilities such as wind turbine towers, substations, and cable lines;

[0016] Step 3: According to the actual geographical environment of the wind farm and the distribution of communication equipment, network nodes are constructed. Each node is a substation, which integrates all the visual information, temperature information, and equipment operation status described in step 1 within the node area;

[0017] Step 4: Formulate the election rules for the master station and substation, adopt the election strategy, and vote to select the master station. Each substation outputs the communication status of the current site, including the number of connected devices, the operating temperature of the communication equipment, the data transmission rate, the communication equipment load, and the host computing power margin. According to the election rules, the substation with the highest score becomes the master station; if the substations of the entire wind power station are divided into multiple groups of substations due to damage to the communication equipment, each group of substations will select its own master station. When the communication is restored, the substation groups will be merged to re-select a new master station;

[0018] Step 5: Formulate mapping rules, define digital feature vectors, and de-dimensionalize the data of each dimension in each node according to the rules, so that each node has a set of digital features; perform independent normalization processing on each element in the digital features, where the digital features include the number of equipment, equipment operating efficiency, ambient temperature, number of people, number of emergency supplies, road code, road smoothness, road risk level, master station substation code, and communication rate;

[0019] Step 6: According to the actual situation in the wind farm, formulate abnormal data indicators and emergency plans of different types of emergency events, where abnormal data indicators include but are not limited to equipment communication success rate, environmental temperature of important facilities, and equipment operation efficiency; emergency event types include equipment failure and natural disasters; collect resource information contained in the indicators in real time, determine the type of emergency event according to the logic of plan formulation, load the initial value of the digital feature vector of the corresponding type of plan, and prepare for subsequent plan evaluation and optimization;

[0020] Step 7: According to the current emergency event type, use genetic algorithm to calculate the best path and establish a mathematical model, including setting decision variables, objective functions, and constraints;

[0021] Furthermore, in step 2, advanced modeling tools and techniques are used to ensure the high accuracy and richness of details of the model; wherein, the high accuracy and richness of details are based on data provided by the Geographic Information System (GIS), accurately restoring the topography around the wind farm, modeling important facilities such as wind towers, substations, and cable lines with the same size as the actual objects, and integrating dynamic elements such as vehicles, personnel, and drones.

[0022] Furthermore, in the steps 3-5, the wind farm is divided into nodes, a node network is constructed, each node is used as a substation, and the master station and the substation are allocated by election; wherein, in the construction of the network node, the data information in the node area is centrally recorded, and the digital features of the node are formed by dedimensionalization and normalization;

[0023] Furthermore, the step 6 mainly includes resource information integration and plan formulation logic, wherein the resource information integration includes material information integration, communication equipment integration, personnel information integration, data synchronization mechanism, and address book management; the plan formulation logic includes risk identification, plan template library, disaster type management, plan evaluation and optimization;

[0024] Furthermore, the decision variables, objective function, and constraints in step 7 are as follows: The decision variable is x ij , is a binary variable, where i and j represent two nodes on the path. If the path from node i to node j is selected, then x ij =1; otherwise, x ij =0; the objective function is divided into two categories. The first category is natural disasters, using the formula Minmize Z1 = Σ i，j (C ij × ij +T ij × ij +R ij ×)x ij , where C ij is the security cost from node i to j, T ij is the estimated time cost, R ij is the risk score of the road. The second category is the equipment failure situation, using the formula Minmize Z2=Σ i，j (D ij × ij +E ij × ij ), where D ij is the distance cost from node i to j, E ij is the estimated economic cost, including maintenance cost and downtime loss; the constraints include flow conservation constraint, resource constraint, and path existence constraint. For the flow conservation constraint, for each node k, except the end point and the starting point, the formula Σ i x ik -Σ j x kj = 0, ensuring that the flow entering any node is equal to the flow leaving the node, resource constraints, using the formula Σ i，j x ij ×M ij ≤M, where M ij is the amount of materials required to go from path i to j, M is the total amount of available materials, and there are path constraints to ensure that each selected path is continuous, that is, the two nodes of the path must be adjacent.

[0025] Furthermore, the genetic algorithm in step 7 randomly generates a set of initial paths, each path representing an evacuation plan. The generation method can be based on a feasible route map by randomly selecting a node sequence between the starting point and the target point. Each individual represents a path selection plan, and new possible solutions are explored through crossover and mutation operations, wherein the crossover operation generates a new path by exchanging part of the node sequence in the two paths. In the case of suitable node encoding, ensure that the generated path is still a valid path. For example, the coordinate encoding at the intersection must be within a certain limit, that is, adjacent sections can only be cross-operated; in order to increase the diversity of solutions, mutation operations can be introduced. Mutation usually randomly changes a part of the path, such as randomly changing a node or exchanging two nodes; and the fitness of each solution is evaluated using the objective function defined above. The higher the fitness, the better the path. Repeat this process until the termination condition is reached, such as the maximum number of iterations or the fitness tends to be stable.

[0026] Compared with the prior art, the present invention has the following beneficial effects:

[0027] (1) Compared with traditional static paper plans, this solution can achieve dynamic update of emergency plan management through digital feature calculation. Emergency plans can be quickly adjusted according to real-time data and changes in emergency situations, effectively avoiding the shortcomings of traditional plans with long update cycles and inability to respond to emergencies in a timely manner;

[0028] (2) Through digital characterization and swarm intelligence optimization algorithms (genetic algorithms), emergency resources can be efficiently dispatched between various management levels, greatly improving the efficiency of emergency response and overcoming the problems of information isolation and untimely resource dispatch among departments in traditional management;

[0029] (3) Unlike traditional decision support that relies on historical data and experience, this solution can provide decision makers with accurate and timely on-site information through real-time data monitoring and intelligent fault diagnosis, ensuring that emergency response decisions are based on actual conditions, thereby greatly improving the timeliness and accuracy of decision-making;

[0030] (4) Compared with the traditional system that is difficult to provide intuitive and comprehensive display of emergency events, this solution enables managers to quickly grasp the progress and overall situation of emergency events through real-time monitoring interface and visual display function, thereby improving the ability to control emergencies. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 Schematic diagram of the method structure provided by the present invention

[0032] Figure 2 Flow chart of the method provided by the present invention

[0033] Figure 3 Network node diagram provided by the present invention

[0034] Figure 4 The path planning index diagram provided by the present invention

[0035] Figure 5 The path planning matrix diagram provided by the present invention

[0036] Figure 6 The path planning diagram output by the genetic algorithm provided by the present invention

[0037] Figure 7 Schematic diagram of the visualization interface provided by the present invention DETAILED DESCRIPTION

[0038] In order to make the solution of the present invention clearer, the present invention will be further described below with reference to the accompanying drawings:

[0039] As the instruction manual Figure 1 The figure shows the structure diagram of the wind farm intelligent emergency dispatch optimization management method based on heuristic algorithm described in the present invention, including information collection and modeling, master-substation collaborative calculation and emergency dispatch algorithm, wherein:

[0040] The information collection and modeling mentioned above, i.e., collecting equipment, personnel, and road information to build a three-dimensional model, can specifically use the three-dimensional model of the earth to obtain the topography, build three-dimensional models of wind turbine towers and substations, and use virtual line technology to map cable lines. In addition, roads are created and numbered based on the longitude and latitude coordinates of equipment and facilities, and cameras are installed in combination with algorithms to assess crowd density, road accessibility, and risk levels. At the same time, ambient temperature and equipment operating conditions are collected through temperature sensors and measuring instruments to provide support for further data analysis and decision-making.

[0041] The master-substation collaborative calculation is to encode the master-substation and define the digital features of the nodes. Specifically, the substation monitoring platform can be used as the master node, and the equipment at each wind turbine site can be used as the subnode. The master node can obtain the status information of the wind turbine equipment, including online, offline, and fault, so as to count the number of wind turbines, towers, and substations. By checking the material flow records, the number of emergency materials can be obtained. The communication equipment is equipped with a temperature sensor to obtain temperature information. The speed and active power parameters of the wind turbine generator are read to obtain the equipment operation efficiency. Considering that the information categories obtained by each subnode are diverse, it is necessary to de-dimensionalize and normalize them to construct a set of digital feature vectors to characterize the status of the current node.

[0042] The emergency dispatch algorithm, i.e., calculating the best behavior path, uses a genetic algorithm to calculate the best path by combining the feature vector information provided by the substation with the global state of the main station. The path here can be a personnel evacuation path or a path with the minimum operation and maintenance cost when equipment fails.

[0043] As the instruction manual Figure 2 The figure shows a method flow chart of the wind farm intelligent emergency dispatch optimization management method based on heuristic algorithm proposed by the present invention, which includes the following steps:

[0044] Step 1: Information data collection, build a three-dimensional model. Use the three-dimensional model of the earth to obtain the topography; use the three-dimensional drawing model to build the wind tower and substation model; use the virtual line to obtain the cable line, create and number the roads according to the longitude and latitude coordinates of the wind turbines and tower equipment; install cameras and combine algorithms to count the number of people, the degree of road traffic and the risk level of the road; use temperature sensors to collect the temperature inside and outside the wind power plant; use measuring instruments to collect the operating status of the equipment;

[0045] Step 2: Based on the collected visual information, use AutoCAD software to render and build a scaled-down model

[0046] Step 3: Build network nodes according to the environment and equipment distribution. Read the longitude and latitude of each wind turbine and tower equipment, take the wind turbine station as the center, and use each wind turbine station as a node to draw a global node distribution map.

[0047] Step 4: Select the master station and substation to assign node tasks. The monitoring center is the master station or the main node, and the wind turbine site is the substation or the subnode. The monitoring center of the master node is responsible for counting the visual information, temperature information, and equipment operation status information of all subnodes. The visual information refers to wind towers, substations, cable lines, crowd numbers, road codes, road smoothness, and road risk levels. The temperature information refers to the indoor temperature and road temperature of the wind turbine site (subnode). The equipment operation status information refers to voltage, current, and active power. The subnode obtains the information of the terminal device through the actual sensors, cameras, and measuring instruments on site.

[0048] Step 5: Define digital feature vectors. The information obtained by each child node is diverse and needs to be de-dimensionalized and normalized to construct a set of digital feature vectors to characterize the state of the current node. Here, each node has the following feature vectors [number of people, road code, road smoothness, road risk level, equipment operating temperature, indoor and outdoor temperature, active power, voltage, current], which can be freely expanded according to actual conditions. In order to briefly explain how to construct a digital feature vector, each node has [equipment operating temperature, number of people, road smoothness] three digital information characterizing the current node. Note that each node has multiple devices and multiple roads. First, define the numerical range for the three features of each node, and divide these ranges into multiple intervals. Each feature will count the number of devices or data points in different intervals according to the results of the interval division to form a feature vector. Feature interval: According to the temperature range Tmin to Tmax, the temperature data is divided into k intervals, according to the number of people at the node Nmin to Nmax, the crowd data is divided into m intervals, according to the road smoothness Cmin to Cmax, it is divided into n intervals, so that each feature will have a corresponding number of intervals; then, for each device in each node, the number of data points falling into the corresponding interval for each temperature, crowd number, and road smoothness is counted. Then, normalization is performed using the formula Among them, h i is the number of point data in a certain interval of a certain feature, N total is the total number of devices in the node. Finally, construct the feature vector of each node. Since the three features are divided into k, m, and n intervals respectively, the feature vector of this node will be a k+m+n dimensional vector in the form of

[0049]

[0050] Step 6: By summarizing historical events, formulate abnormal data indicators and emergency plans for different types of emergency events. At the same time, each node feature vector obtained in step 5 is converted into a path planning cost vector [CRTDE] through expert experience, where C represents safety cost, R represents road risk score, T represents estimated time cost, D represents distance cost, and E represents economic cost. These cost feature vectors are used for subsequent path calculations. Among them, safety cost can be extracted from indoor and outdoor temperature and equipment operation status elements of node feature vectors, road risk score can be extracted from the number of people in node feature vectors combined with earthquake statistics, estimated time cost can be obtained from commuting time records, distance cost can be extracted from road coding of node feature vectors, economic cost can be extracted from active power and equipment maintenance cost of node feature vectors, communication success rate in abnormal data indicators can be extracted from online and offline equipment information, natural disaster information can be extracted from temperature information of historical fire situations, and emergency plans can be formulated from expert experience to formulate regular evacuation start and end points.

[0051] Step 7: Apply genetic algorithm to calculate the best path. To simplify the explanation, we assume that the network has 30 nodes. Figure 3 As shown in the figure, corresponding to 30 wind turbine sub-nodes, combined with the following constraints, the optimal path is calculated by genetic algorithm when the input is [2,9,14,21,28] and [6,12,18,24,30]. These two arrays correspond to the path starting point array and the path ending point array, respectively, as shown in Figure 4 As shown in , the fitness of the final 5 best routes is higher than the fitness of other paths in the iterative process, such as Figure 5 As shown, each row of the final output matrix represents a path:

[0052] 1) There are two input arrays with a dimension of 5. The first is the path starting point array. There are as many paths to be planned as there are non-zero elements. The second array is the path ending point. The ending point of each planned path must be one of the non-zero elements in the second array. The input array here comes from the initial plan in step 6.

[0053] 2) The entire network is a 5*6 node matrix. Each node has its own adjacent node array. Each element in this array has a eigenvector [CRTDE]. For example, the adjacent node array of node 3 is [2,4,9]. Element 2 in this array has a eigenvector [2,4,1,2,4]. This means that the comprehensive cost of node 3 reaching node 2 is 2+4+1+2+4. Similarly, elements 4 and 9 in this array also have corresponding eigenvectors.

[0054] 3) Path planning is done using a genetic algorithm. All operations in the algorithm, such as mutation and crossover, can only be selected in the adjacent node array of the previous node. For example, there is a path with a starting point of 9 and an end point array of [12, 13, 19, 0, 0]. The initial path can be to only consider the number of nodes on the path, such as 9, 8, 7, 13, 0, 0 or 9, 10, 11, 12, 0, 0. For example, 9, 8, 7, 13, 0, 0 is taken, and then genetic mutation occurs, and node 8 is no longer needed. Since node 8 is the second element and its previous element is 9, then node 8 is the second element. Point 8 can only mutate to other elements in the adjacent node array of node 9. Here, the adjacent node array corresponding to node 9 is [3,8,10,15]. For example, if it mutates to 10, please note that since the second element of the path is now 10, and 10 is incompatible with the original path, because there is no 7 in the adjacent element array of 10, the third element must also mutate and can only be selected from the adjacent node array of 10, that is, [4,9,11,16], and so on. In other words, if the current node wants to mutate or crossover, it is constrained by the adjacent node array of the previous node;

[0055] 4) If the node is the starting point of the path, then the adjacent node arrays of all nodes should remove all the starting nodes of the path from the array. For example, in the example given in condition 3, the starting node is 9, so the adjacent node array of node 3 is [2,4,9], which should be changed to [2,4];

[0056] 5) All operations of the genetic algorithm can only be selected from adjacent nodes to ensure the continuity of the path;

[0057] 6) Because the output is a 5*6 matrix, each row represents a path, so in the genetic algorithm, the starting point of the path cannot be changed, which is equivalent to the first column of the matrix elements being unchanged. Of course, because there are 5 rows, if the input node starting vector has less than 5 non-zero elements, then the first column of the matrix will be filled with zeros. Similarly, because the output is a 5*6 matrix, it must be ensured that the path must be completed within 6 nodes, that is, the last non-zero element of each row must be one of the elements in the second array of the input, that is, the path end node array;

[0058] 7) The eigenvector here is mainly used to calculate the fitness function. The higher the function value, the better. Therefore, there is such a phenomenon. Sometimes the path is short, such as 9, 10, 11, 12, 0, 0, and the other path is 9, 8, 7, 13, 19, 0. Although the latter path has more nodes, due to the different eigenvectors, the path with more nodes may have a smaller cost. Here we choose the path with the smallest cost. That is to say, when initially generating the path, we can only consider the number of nodes, but ultimately consider the path cost.

Claims

1. A wind farm intelligent emergency dispatch optimization management method based on a heuristic algorithm, characterized in that: The following steps are involved: Step 1: Use image acquisition equipment to collect visual information, including topography, wind turbine towers, substations, cable lines, crowd numbers, road codes, road accessibility, and road risk levels; use temperature acquisition equipment to collect temperature information, including the temperature values ​​inside and outside the wind turbine station; use measuring instruments to collect equipment operating status, including voltage, current, and power; Step 2: Based on the collected visual information, build a 3D model of important facilities such as wind turbine towers, substations, and cable lines; Step 3: According to the actual geographical environment of the wind farm and the distribution of communication equipment, network nodes are constructed. Each node is a substation, which integrates all the visual information, temperature information, and equipment operation status described in step 1 within the node area; Step 4: Formulate the election rules for the master station and substation, adopt the election strategy, and vote to select the master station. Each substation outputs the communication status of the current site, including the number of connected devices, the operating temperature of the communication equipment, the data transmission rate, the communication equipment load, and the host computing power margin. According to the election rules, the substation with the highest score becomes the master station; if the substations of the entire wind power station are divided into multiple groups of substations due to damage to the communication equipment, each group of substations will select its own master station. When the communication is restored, the substation groups will be merged to re-select a new master station; Step 5: Formulate mapping rules, define digital feature vectors, and de-dimensionalize the data of each dimension in each node according to the rules, so that each node has a set of digital features; perform independent normalization processing on each element in the digital features, where the digital features include the number of equipment, equipment operating efficiency, temperature inside and outside the wind turbine station, number of people, number of emergency supplies, road code, road smoothness, road risk level, master station substation code, and communication rate; Step 6: According to the actual situation in the wind farm, formulate abnormal data indicators and emergency plans of different types of emergency events, where abnormal data indicators include but are not limited to equipment communication success rate, environmental temperature of important facilities, and equipment operation efficiency; emergency event types include equipment failure and natural disasters; collect resource information contained in the indicators in real time, determine the type of emergency event according to the logic of plan formulation, load the initial value of the digital feature vector of the corresponding type of plan, and prepare for subsequent plan evaluation and optimization; Step 7: According to the current emergency event type, use genetic algorithm to calculate the best path and establish a mathematical model, including setting decision variables, objective function, and constraints.

2. The method for optimizing management of wind farm intelligent emergency dispatch based on heuristic algorithm according to claim 1 is characterized in that: Step 2, using advanced modeling tools and techniques to ensure high accuracy and rich details of the model; Among them, the high precision and richness of details are based on data provided by the Geographic Information System (GIS), which accurately restores the terrain around the wind farm, models important facilities such as wind towers, substations, and cable lines with the same size as the actual objects, and integrates dynamic elements such as vehicles, personnel, and drones.

3. The wind farm intelligent emergency dispatch optimization management method based on heuristic algorithm according to claim 1 is characterized in that: The steps 3-5 divide the wind farm into nodes, construct a node network, each node as a substation, and allocate the master station and substation through election; Among them, the construction of network nodes records the data information in the node area in a centralized manner, and forms the digital characteristics of the nodes through dedimensionalization and normalization.

4. The method for optimizing management of wind farm intelligent emergency dispatch based on heuristic algorithm according to claim 1 is characterized in that: The step 6 mainly includes resource information integration and plan formulation logic; Among them, the resource information integration includes material information integration, communication equipment integration, personnel information integration, data synchronization mechanism, and address book management; the plan formulation logic includes risk identification, plan template library, disaster type management, plan evaluation and optimization.

5. The method for optimizing management of wind farm intelligent emergency dispatch based on heuristic algorithm according to claim 1 is characterized in that: The plan evaluation and optimization mentioned in step 6 adopts a genetic algorithm.

6. The wind farm intelligent emergency dispatch optimization management method based on heuristic algorithm according to claim 6 is characterized in that: The decision variables, objective function, and constraints in step 7 are as follows: The decision variable is x ij , is a binary variable, where i and j represent two nodes on the path; if the path from node i to node j is selected, then x ij =1; Otherwise, x ij =0; The objective function is divided into two categories. The first category is natural disasters, using the formula Minimize Z1 = Σ i，j (C ij × ij +T ij × ij +R ij × ij ), where C ij is the security cost from node i to j, T ij is the estimated time cost, R ij is the risk score of the road. The second category is the equipment failure situation, using the formula Minimize Z2=∑ i，j (D ij × ij +E ij × ij ), where D ij is the distance cost from node i to j, E ij is the estimated economic cost, including maintenance cost and downtime loss; the constraints include flow conservation constraint, resource constraint, and path existence constraint. For the flow conservation constraint, for each node k, except the end point and the starting point, the formula ∑ i x ik -∑ j x kj = 0, ensuring that the flow entering any node is equal to the flow leaving the node, resource constraints, using the formula ∑ i，j x ij ×M ij ≤M, where M ij is the amount of materials required to go from path i to j, M is the total amount of available materials, and there are constraints on the path to ensure that each selected path is continuous, that is, the two nodes of the path must be adjacent.

7. The method for optimizing management of wind farm intelligent emergency dispatch based on heuristic algorithm according to claim 6 is characterized in that: The genetic algorithm in step 7 randomly generates a set of initial paths, each path representing an evacuation plan; The generation method is based on a feasible route map by randomly selecting a node sequence between the starting point and the target point; each individual represents a path selection plan, and new possible solutions are explored through crossover and mutation operations, where the crossover operation generates a new path by exchanging part of the node sequence in the two paths; when node encoding is suitable, ensure that the generated path is still a valid path; the coordinate encoding at the intersection must be within a certain limit, that is, adjacent sections can only be crossed; in order to increase the diversity of solutions, mutation operations are introduced; mutations usually randomly change a part of the path, randomly change a node or exchange two nodes; and use the objective function defined above to evaluate the fitness of each solution. The higher the fitness, the better the path. Repeat this process until the termination condition is reached, such as the maximum number of iterations or the fitness tends to be stable.

8. A computer-readable storage medium, characterized in that: The computing program for implementing the management method as claimed in claims 1-7 is deployed on a main server selected in the wind farm station, and when the server executes, it implements all the steps as claimed in claims 1-7.

Images