A wind turbine group operation and maintenance method considering multiple time windows and intelligent partitioning and related device

Abstract

This invention provides a wind turbine cluster operation and maintenance method and related apparatus considering multiple time windows and intelligent partitioning, belonging to the field of offshore wind turbine cluster operation and maintenance strategies. The invention first determines the opportunity maintenance interval for offshore wind turbines and determines time windows considering the safety and economy of maintenance tasks, thereby calculating multi-level operation and maintenance time windows. With the goal of reducing the total cost within the maintenance cycle, a dynamic overall maintenance model for offshore wind turbines and a dynamic maintenance sub-model for each maintenance vessel are constructed. The turbines to be maintained are partitioned, and maintenance tasks are assigned to each vessel. By solving the dynamic maintenance sub-models of each partition, the operation and maintenance strategy for the wind turbines is obtained. This invention improves strategy flexibility by solving for the daily optimal maintenance strategy and dynamically updating the strategy by combining maintenance information and environmental change information. Simultaneously, the intelligent partitioning strategy decomposes the overall maintenance model into multiple sub-models, effectively reducing model complexity and improving the efficiency of solving the maintenance strategy.

Description

Technical Field

[0001] This invention belongs to the field of offshore wind turbine cluster operation and maintenance strategy technology, specifically involving a wind turbine cluster operation and maintenance method and related devices that consider multiple time windows and intelligent partitioning. Background Technology

[0002] Due to the harsh marine environment, wind turbines are constantly subjected to impacts such as typhoons, seawater corrosion, and lightning strikes, significantly increasing the operation and maintenance costs of offshore wind power. According to relevant data, the operation and maintenance costs of offshore wind turbines severely impact the economic benefits of offshore wind farms. Therefore, proposing a scientific and economical operation and maintenance strategy for offshore wind turbines is of great significance for improving the economics of offshore wind power and promoting its further development.

[0003] Offshore wind power operation and maintenance (O&M) is a complex dynamic optimization problem that requires consideration of factors such as component condition, maintenance environment, maintenance costs, and maintenance resources. Preventive opportunistic maintenance (POM) strategy refers to providing other wind turbines within the opportunistic maintenance window with a maintenance opportunity when a turbine that has reached a reliability threshold needs maintenance or repair, in order to reduce the number of maintenance trips and costs. This can significantly reduce O&M costs. However, while there has been considerable research on offshore wind power O&M strategies, the following challenges remain: 1) Offshore wind turbine maintenance is affected by multiple factors such as component reliability, sea conditions, and weather, making it difficult to construct a scientifically reasonable O&M time window to achieve safe and economical maintenance; 2) The marine environment is highly uncertain, and turbine failures can be sudden, making it difficult to dynamically update maintenance strategies to flexibly respond to environmental changes and sudden turbine failures; 3) Large-scale offshore wind farms are massive, and the maintenance strategy models are highly complex, making it difficult to reduce the complexity of the maintenance models and improve the solution speed. Summary of the Invention

[0004] In order to at least partially solve one of the technical problems existing in the prior art, the present invention aims to provide a wind turbine group operation and maintenance method and related device that considers multiple time windows and intelligent partitioning.

[0005] To achieve the above objectives, the technical solution provided by the present invention is as follows:

[0006] In a first aspect, the present invention provides a wind turbine group operation and maintenance method that considers multiple time windows and intelligent partitioning, comprising the following steps:

[0007] Determine the opportunity maintenance interval for each offshore wind turbine;

[0008] Determine the time window for each offshore wind turbine to consider the safety of maintenance tasks and the time window to consider the economics of offshore wind power. Combine the opportunity maintenance interval and the time corresponding to the determined time window to obtain the multi-level operation and maintenance time window for each offshore wind turbine.

[0009] With the goal of minimizing the total cost of each maintenance task within the maintenance cycle, a dynamic maintenance large model for offshore wind turbines is established, and a dynamic maintenance small model for each maintenance vessel is obtained based on the dynamic maintenance large model.

[0010] For the units to be maintained that fall within the corresponding multi-level maintenance time window on that day, the units to be maintained are divided into zones based on the number of maintenance vessels dispatched that day, and a maintenance task for each zone is assigned to each maintenance vessel.

[0011] The dynamic maintenance mini-model corresponding to each partition maintenance task is solved to obtain the operation and maintenance strategy of the unit to be maintained.

[0012] Furthermore, in the step of determining the opportunity maintenance interval for each offshore wind turbine, the opportunity maintenance interval for the offshore wind turbine is obtained by utilizing the three-parameter Weibull distribution and reliability analysis.

[0013] The opportunity maintenance interval for offshore wind turbines is obtained using a three-parameter Weibull distribution and reliability analysis, including:

[0014] Based on the three-parameter Weibull distribution and reliability analysis, the historical fault data of each offshore wind turbine were fitted to obtain the failure probability density function, failure probability function, reliability function and failure rate function of each wind turbine component;

[0015] Set the opportunity maintenance reliability threshold and the preventive maintenance reliability threshold, and substitute them into the reliability function. Solve the failure probability density function, failure probability function, reliability function and failure rate function simultaneously to obtain the opportunity maintenance interval corresponding to the opportunity maintenance reliability threshold and the preventive maintenance reliability threshold.

[0016] Furthermore, the steps of determining the time window for each offshore wind turbine considering the safety of maintenance tasks and the time window considering the economics of offshore wind power, and combining the opportunistic maintenance interval with the time corresponding to the determined time window to obtain the multi-level operation and maintenance time window for each offshore wind turbine, include:

[0017] The time window that considers the security of the maintenance task is determined by using the security of the maintenance task as a constraint, and is denoted as the first-level time window T. d ;

[0018] The opportunity maintenance interval is denoted as the secondary time window [T]. o T p ], T o To maintain the initial time of the opportunity, Tp To maintain the opportunity at the end of the time;

[0019] The time window for considering the economics of offshore wind power is determined by using power generation loss as a constraint, and denoted as T. loss ;

[0020] Let T be the maintenance time corresponding to the initial reliability of the offshore wind turbine. c If the maintenance time T c If it is within the second-level time window, then the multi-level operation and maintenance time window is... If maintenance time T c If it is not within the second-level time window, then the multi-level operation and maintenance time window is... .

[0021] Furthermore, the expressions for dynamically maintaining the large model and dynamically maintaining the small model are as follows:

[0022] The objective function for dynamically maintaining the large model is as follows:

[0023] ;

[0024] In the formula, For total cost, For the cost of travel, To incur penalties, To maintain costs, For vessel leasing costs, This is due to the loss of power generation.

[0025] The objective function for dynamically maintaining the small model is as follows:

[0026] ;

[0027] In the formula, , , and Let represent the total cost, navigation cost, penalty cost, and power generation loss of the i-th partition on that day, respectively;

[0028] The constraints of the model are as follows:

[0029] Reliability constraints:

[0030] ;

[0031] In the formula, This represents the preventive maintenance reliability threshold for the i-th unit. This represents the opportunity maintenance reliability threshold for the i-th generating unit;

[0032] Daily working hours constraints:

[0033] ;

[0034] In the formula, This represents the maintenance time for the i-th generating unit. This indicates the total sailing time of the maintenance vessel, with 12 hours representing the daily working time.

[0035] Furthermore, in the step of dividing the units to be maintained into zones based on the number of maintenance vessels dispatched that day, for units falling within the corresponding multi-level maintenance time window, the following steps are included:

[0036] Based on the multi-level maintenance time window, the units to be maintained are divided into alarm maintenance units and opportunity maintenance units; alarm maintenance units are units whose current date is the last day of the corresponding multi-level maintenance time window; opportunity maintenance units are all units in the units to be maintained except alarm maintenance units;

[0037] Based on the number of maintenance vessels dispatched on that day, a clustering algorithm is used to cluster the maintenance units to be maintained, and each resulting cluster is used as a partition. If there are no alarm maintenance units on that day, the opportunity maintenance units are directly clustered according to the number of vessels dispatched. If there are alarm maintenance units on that day, each alarm maintenance unit is used as a cluster center and then clustered according to the number of vessels dispatched.

[0038] Furthermore, the step of clustering the units to be maintained using a clustering algorithm based on the number of maintenance vessels dispatched that day includes:

[0039] If there are no alarm maintenance units on that day, the number of ships dispatched will be used as the cluster number, and clustering will be achieved using a clustering algorithm.

[0040] If the number of alarm maintenance units on the same day is equal to the number of ships dispatched, then the number of ships dispatched is used as the cluster number, and the alarm maintenance units on the same day are used as the initial cluster centers. Clustering is achieved using a clustering algorithm.

[0041] If the number of alarm maintenance units on a given day is not equal to the number of ships dispatched, then the number of alarm maintenance units is used as the cluster number, and the alarm maintenance units on that day are used as the initial cluster centers. The initial clustering results are obtained using a clustering algorithm. Based on the principle of maximizing the inter-cluster distance, several clusters in the initial clustering results are aggregated or decomposed according to the number of ships dispatched.

[0042] Furthermore, in the step of solving the dynamic maintenance mini-model corresponding to each partition maintenance task to obtain the operation and maintenance strategy of the unit to be maintained, an improved genetic algorithm is used for the solution. The solution process includes the following steps:

[0043] Initialize a 0-1 population; each particle in the 0-1 population represents an operation and maintenance scheme. Each operation and maintenance scheme satisfies the daily working time constraint and includes a decision variable indicating whether to maintain the units to be maintained in its respective partition. Here, a decision variable of 0 indicates no maintenance and a decision variable of 1 indicates maintenance.

[0044] A partial matching crossover method and an adaptive mutation method were used to perform crossover and mutation operations on the 0-1 population.

[0045] The selection operation is based on the fitness value, selecting particles with a fitness value exceeding the set value from the current population to enter the next generation of the population;

[0046] Repeatedly perform crossover and selection operations on the next generation population until the preset termination conditions are met, and determine the operation and maintenance scheme corresponding to the particle with the highest fitness in the final population as the operation and maintenance strategy of the unit to be maintained in the corresponding partition.

[0047] Secondly, the present invention provides a wind turbine group operation and maintenance device that considers multiple time windows and intelligent zoning, comprising:

[0048] The multi-time window determination module is used to determine the opportunity maintenance interval for each offshore wind turbine; it is also used to determine the time window for each offshore wind turbine considering the safety of maintenance tasks and the time window considering the economics of offshore wind power, and to obtain the multi-level operation and maintenance time window for each offshore wind turbine by combining the opportunity maintenance interval and the time corresponding to the determined time window.

[0049] The model building module is used to establish a dynamic maintenance large model of offshore wind turbines with the goal of minimizing the total cost of each maintenance task within the maintenance cycle, and to decompose the dynamic maintenance small model of each maintenance vessel based on the dynamic maintenance large model.

[0050] The intelligent partitioning module is used to partition the units to be maintained according to the number of maintenance vessels dispatched on the same day, based on the corresponding multi-level maintenance time window, and assign a maintenance task to each maintenance vessel for that partition.

[0051] The model solving module is used to solve the dynamic maintenance mini-model corresponding to each partition maintenance task to obtain the operation and maintenance strategy of the unit to be maintained.

[0052] Furthermore, in the step of determining the opportunity maintenance interval for each offshore wind turbine, the opportunity maintenance interval for the offshore wind turbine is obtained by utilizing the three-parameter Weibull distribution and reliability analysis.

[0053] The opportunity maintenance interval for offshore wind turbines is obtained using a three-parameter Weibull distribution and reliability analysis, including:

[0054] Based on the three-parameter Weibull distribution and reliability analysis, the historical fault data of each offshore wind turbine were fitted to obtain the failure probability density function, failure probability function, reliability function and failure rate function of each wind turbine component;

[0055] Set the opportunity maintenance reliability threshold and the preventive maintenance reliability threshold, and substitute them into the reliability function. Solve the failure probability density function, failure probability function, reliability function and failure rate function simultaneously to obtain the opportunity maintenance interval corresponding to the opportunity maintenance reliability threshold and the preventive maintenance reliability threshold.

[0056] Furthermore, the steps of determining the time window for each offshore wind turbine considering the safety of maintenance tasks and the time window considering the economics of offshore wind power, and combining the opportunistic maintenance interval with the time corresponding to the determined time window to obtain the multi-level operation and maintenance time window for each offshore wind turbine, include:

[0057] The time window that considers the security of the maintenance task is determined by using the security of the maintenance task as a constraint, and is denoted as the first-level time window T. d ;

[0058] The opportunity maintenance interval is denoted as the secondary time window [T]. o T p ], T o To maintain the initial time of the opportunity, T p To maintain the opportunity at the end of the time;

[0059] The time window for considering the economics of offshore wind power is determined by using power generation loss as a constraint, and denoted as T. loss ;

[0060] Let T be the maintenance time corresponding to the initial reliability of the offshore wind turbine. c If the maintenance time T c If it is within the second-level time window, then the multi-level operation and maintenance time window is... If maintenance time T c If it is not within the second-level time window, then the multi-level operation and maintenance time window is... .

[0061] Furthermore, the expressions for dynamically maintaining the large model and dynamically maintaining the small model are as follows:

[0062] The objective function for dynamically maintaining the large model is as follows:

[0063] ;

[0064] In the formula, For total cost, For the cost of travel, To incur penalties, To maintain costs, For vessel leasing costs, This is due to the loss of power generation.

[0065] The objective function for dynamically maintaining the small model is as follows:

[0066] ;

[0067] In the formula, , , and Let represent the total cost, navigation cost, penalty cost, and power generation loss of the i-th partition on that day, respectively;

[0068] The constraints of the model are as follows:

[0069] Reliability constraints:

[0070] ;

[0071] In the formula, This represents the preventive maintenance reliability threshold for the i-th unit. This represents the opportunity maintenance reliability threshold for the i-th generating unit;

[0072] Daily working hours constraints:

[0073] ;

[0074] In the formula, This represents the maintenance time for the i-th generating unit. This indicates the total sailing time of the maintenance vessel, with 12 hours representing the daily working time.

[0075] Furthermore, in the step of dividing the units to be maintained into zones based on the number of maintenance vessels dispatched that day, for units falling within the corresponding multi-level maintenance time window, the following steps are included:

[0076] Based on the multi-level maintenance time window, the units to be maintained are divided into alarm maintenance units and opportunity maintenance units; alarm maintenance units are units whose current date is the last day of the corresponding multi-level maintenance time window; opportunity maintenance units are all units in the units to be maintained except alarm maintenance units;

[0077] Based on the number of maintenance vessels dispatched on that day, a clustering algorithm is used to cluster the maintenance units to be maintained, and each resulting cluster is used as a partition. If there are no alarm maintenance units on that day, the opportunity maintenance units are directly clustered according to the number of vessels dispatched. If there are alarm maintenance units on that day, each alarm maintenance unit is used as a cluster center and then clustered according to the number of vessels dispatched.

[0078] Furthermore, the step of clustering the units to be maintained using a clustering algorithm based on the number of maintenance vessels dispatched that day includes:

[0079] If there are no alarm maintenance units on that day, the number of ships dispatched will be used as the cluster number, and clustering will be achieved using a clustering algorithm.

[0080] If the number of alarm maintenance units on the same day is equal to the number of ships dispatched, then the number of ships dispatched is used as the cluster number, and the alarm maintenance units on the same day are used as the initial cluster centers. Clustering is achieved using a clustering algorithm.

[0081] If the number of alarm maintenance units on a given day is not equal to the number of ships dispatched, then the number of alarm maintenance units is used as the cluster number, and the alarm maintenance units on that day are used as the initial cluster centers. The initial clustering results are obtained using a clustering algorithm. Based on the principle of maximizing the inter-cluster distance, several clusters in the initial clustering results are aggregated or decomposed according to the number of ships dispatched.

[0082] Furthermore, in the step of solving the dynamic maintenance mini-model corresponding to each partition maintenance task to obtain the operation and maintenance strategy of the unit to be maintained, an improved genetic algorithm is used for the solution. The solution process includes the following steps:

[0083] Initialize a 0-1 population; each particle in the 0-1 population represents an operation and maintenance scheme. Each operation and maintenance scheme satisfies the daily working time constraint and includes a decision variable indicating whether to maintain the units to be maintained in its respective partition. Here, a decision variable of 0 indicates no maintenance and a decision variable of 1 indicates maintenance.

[0084] A partial matching crossover method and an adaptive mutation method were used to perform crossover and mutation operations on the 0-1 population.

[0085] The selection operation is based on the fitness value, selecting particles with a fitness value exceeding the set value from the current population to enter the next generation of the population;

[0086] Repeatedly perform crossover and selection operations on the next generation population until the preset termination conditions are met, and determine the operation and maintenance scheme corresponding to the particle with the highest fitness in the final population as the operation and maintenance strategy of the unit to be maintained in the corresponding partition.

[0087] Thirdly, the present invention provides a computer device, the device including a processor and a memory:

[0088] The memory is used to store computer programs and send the instructions of the computer programs to the processor;

[0089] The processor executes the following steps according to the instructions of the computer program:

[0090] Determine the opportunity maintenance interval for each offshore wind turbine;

[0091] Determine the time window for each offshore wind turbine to consider the safety of maintenance tasks and the time window to consider the economics of offshore wind power. Combine the opportunity maintenance interval and the time corresponding to the determined time window to obtain the multi-level operation and maintenance time window for each offshore wind turbine.

[0092] With the goal of minimizing the total cost of each maintenance task within the maintenance cycle, a dynamic maintenance large model for offshore wind turbines is established, and a dynamic maintenance small model for each maintenance vessel is obtained based on the dynamic maintenance large model.

[0093] For the units to be maintained that fall within the corresponding multi-level maintenance time window on that day, the units to be maintained are divided into zones based on the number of maintenance vessels dispatched that day, and a maintenance task for each zone is assigned to each maintenance vessel.

[0094] The dynamic maintenance mini-model corresponding to each partition maintenance task is solved to obtain the operation and maintenance strategy of the unit to be maintained.

[0095] Fourthly, the present invention provides a computer-readable storage medium on which a computer program is stored, and when executed by a processor, the computer program performs the following steps:

[0096] Determine the opportunity maintenance interval for each offshore wind turbine;

[0097] Determine the time window for each offshore wind turbine to consider the safety of maintenance tasks and the time window to consider the economics of offshore wind power. Combine the opportunity maintenance interval and the time corresponding to the determined time window to obtain the multi-level operation and maintenance time window for each offshore wind turbine.

[0098] With the goal of minimizing the total cost of each maintenance task within the maintenance cycle, a dynamic maintenance large model for offshore wind turbines is established, and a dynamic maintenance small model for each maintenance vessel is obtained based on the dynamic maintenance large model.

[0099] For the units to be maintained that fall within the corresponding multi-level maintenance time window on that day, the units to be maintained are divided into zones based on the number of maintenance vessels dispatched that day, and a maintenance task for each zone is assigned to each maintenance vessel.

[0100] The dynamic maintenance mini-model corresponding to each partition maintenance task is solved to obtain the operation and maintenance strategy of the unit to be maintained.

[0101] Fifthly, the present invention also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0102] Determine the opportunity maintenance interval for each offshore wind turbine;

[0103] Determine the time window for each offshore wind turbine to consider the safety of maintenance tasks and the time window to consider the economics of offshore wind power. Combine the opportunity maintenance interval and the time corresponding to the determined time window to obtain the multi-level operation and maintenance time window for each offshore wind turbine.

[0104] With the goal of minimizing the total cost of each maintenance task within the maintenance cycle, a dynamic maintenance large model for offshore wind turbines is established, and a dynamic maintenance small model for each maintenance vessel is obtained based on the dynamic maintenance large model.

[0105] For the units to be maintained that fall within the corresponding multi-level maintenance time window on that day, the units to be maintained are divided into zones based on the number of maintenance vessels dispatched that day, and a maintenance task for each zone is assigned to each maintenance vessel.

[0106] The dynamic maintenance mini-model corresponding to each partition maintenance task is solved to obtain the operation and maintenance strategy of the unit to be maintained.

[0107] In summary, this invention provides a wind turbine cluster operation and maintenance method and related apparatus that considers multiple time windows and intelligent partitioning. The method first determines the opportunistic maintenance interval for each offshore wind turbine, and then determines the time window for each turbine considering maintenance task safety and offshore wind power economic benefits. By combining the opportunistic maintenance interval and the determined time window, a multi-level operation and maintenance time window is calculated for each offshore wind turbine. With minimizing the total cost of all maintenance tasks within the maintenance cycle as the optimization objective, a dynamic overall maintenance model for offshore wind turbines is constructed, and based on this model, a dynamic maintenance sub-model for each maintenance vessel is further refined. For turbines within the corresponding multi-level operation and maintenance time window, partitioning is performed based on the number of maintenance vessels dispatched that day, and a maintenance task within a partition is assigned to each maintenance vessel. By solving the dynamic maintenance sub-model corresponding to the maintenance tasks in each partition, the operation and maintenance strategy for the turbines to be maintained is obtained. This invention improves the flexibility of the strategy by independently solving the optimal maintenance strategy each day and dynamically updating the maintenance strategy by combining turbine maintenance information and environmental change information. Meanwhile, by using an intelligent partitioning strategy to decompose a single overall maintenance model into multiple sub-models, the complexity of the model is effectively reduced and the solution efficiency of the maintenance strategy is improved. Attached Figure Description

[0108] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0109] Figure 1A flowchart of a wind turbine group operation and maintenance method considering multiple time windows and intelligent partitioning provided in an embodiment of the present invention;

[0110] Figure 2 This is a schematic diagram of the preventive opportunity maintenance strategy process provided in an embodiment of the present invention;

[0111] Figure 3 This is a schematic diagram of the reliability curve of a wind turbine component provided in an embodiment of the present invention;

[0112] Figure 4 This is a schematic diagram illustrating the dynamic update of multi-level operation and maintenance time windows provided in an embodiment of the present invention;

[0113] Figure 5 A flowchart of the intelligent partitioning strategy provided in an embodiment of the present invention;

[0114] Figure 6 This is a schematic diagram of partitioned clustering aggregation provided in an embodiment of the present invention;

[0115] Figure 7 This is a schematic diagram of the location of offshore wind turbines provided in an embodiment of the present invention;

[0116] Figure 8 A Gantt chart for maintenance strategies that only consider first- and second-level time windows, provided for embodiments of the present invention;

[0117] Figure 9 A Gantt chart of maintenance strategy considering multi-level operation and maintenance time windows for wind turbine components provided in this embodiment of the invention;

[0118] Figure 10 A block diagram of a wind turbine group operation and maintenance device considering multiple time windows and intelligent partitioning is provided for an embodiment of the present invention;

[0119] Figure 11 This is a block diagram of a computer device provided in an embodiment of the present invention. Detailed Implementation

[0120] To make the objectives, features, and advantages of this invention more apparent and understandable, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0121] In the description of this invention, it should be understood that the orientation descriptions, such as up, down, front, back, left, right, etc., are based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting this invention.

[0122] In the description of this invention, "several" means one or more, "more than" means two or more, "greater than," "less than," and "exceeding" are understood to exclude the stated number, while "above," "below," and "within" are understood to include the stated number. The use of "first" and "second" in the description is merely for distinguishing technical features and should not be construed as indicating or implying relative importance, or implicitly indicating the number of indicated technical features, or implicitly indicating the order of the indicated technical features.

[0123] Furthermore, in the description of this invention, unless otherwise stated, "multiple" means two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. The character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0124] In the description of this invention, unless otherwise explicitly defined, terms such as "setting," "installing," and "connecting" should be interpreted broadly, and those skilled in the art can reasonably determine the specific meaning of the above terms in this invention in conjunction with the specific content of the technical solution.

[0125] First, some of the terms involved in this invention will be explained as follows:

[0126] K-means: K-cluster average centroid clustering is a classic unsupervised learning algorithm, mainly used to divide a dataset into K clusters (categories), so that data points within the same cluster are as similar as possible, and data points in different clusters are as different as possible.

[0127] The Traveling Salesman Problem (TSP) aims to find the shortest path that allows a traveling salesman to visit a set of cities exactly once and eventually return to the starting point. This is a classic NP-hard problem, meaning that the computational complexity increases dramatically as the number of cities increases.

[0128] PMX: Partially Mapped Crossover. PMX is a crossover method specifically designed for permutation encoding, aiming to generate valid permutations of offspring while preserving some structural information of the parent generation. It ensures the validity of the offspring permutations by exchanging parts of the parent generation and resolving conflicts through mapping relationships.

[0129] This invention provides a wind turbine cluster operation and maintenance method and related apparatus that considers multiple time windows and intelligent partitioning, to overcome the following problems in the prior art: 1) The maintenance of offshore wind turbines is affected by multiple factors such as component reliability, sea conditions and weather, making it difficult to construct a scientific and reasonable operation and maintenance time window to achieve safe and economical maintenance of offshore wind turbines; 2) The marine environment has great uncertainty, and wind turbine failures also have a certain degree of suddenness, making it difficult to dynamically update maintenance strategies to flexibly respond to environmental changes and sudden wind turbine failures; 3) Large-scale offshore wind farms are large in scale, and the maintenance strategy model is highly complex, making it difficult to reduce the complexity of the maintenance model and improve the solution speed of the model.

[0130] Please see Figure 1 This invention first provides a wind turbine group operation and maintenance method that considers multiple time windows and intelligent zoning, including the following steps:

[0131] S1: Determine the opportunity maintenance interval for each offshore wind turbine.

[0132] It should be noted that the opportunistic maintenance interval refers to a relatively suitable time period determined by comprehensively considering factors such as equipment condition monitoring data, historical maintenance records, and operating environment during the operation of offshore wind turbines. Maintenance within this time period can minimize the probability of equipment failure and reduce maintenance costs. For example, by real-time monitoring and analysis of vibration, temperature, and other parameters of key components of the wind turbine (such as gearboxes and generators), predicting that a certain component may experience performance degradation or failure within a certain future timeframe, this timeframe can be considered the opportunistic maintenance interval.

[0133] S2: Determine the time window for each offshore wind turbine to consider the safety of maintenance tasks and the time window to consider the economics of offshore wind power, and combine the opportunity maintenance interval and the time corresponding to the determined time window to obtain the multi-level operation and maintenance time window for each offshore wind turbine.

[0134] It should be noted that because the maintenance of offshore wind turbines is significantly affected by the marine environment (such as wind, waves, and tides), maintenance operations must be carried out within specific time periods to ensure the safety of maintenance personnel and equipment. This time period is the window for considering the safety of the maintenance task. For example, performing maintenance work during periods of calm and clear weather can reduce safety risks; this period of calm and good weather constitutes the safety window.

[0135] The revenue generated by offshore wind farms is closely related to the operating time of the wind turbines. To maximize the economic benefits of a wind farm, maintenance should be performed during periods when the wind turbines have high power generation efficiency and relatively low power generation costs. This time window is the window for considering the economic viability of offshore wind power. For example, during periods of abundant wind resources, wind turbines generate a large amount of electricity, and maintenance at this time would result in significant power loss. Therefore, maintenance should be avoided during these periods and should be performed during times when wind speeds are relatively weak and power generation efficiency is relatively low to minimize economic losses.

[0136] The multi-level operation and maintenance time window is the final time window obtained by comprehensively considering the opportunity maintenance interval, the maintenance task safety time window, and the offshore wind power economic time window. It includes time constraints and requirements in multiple dimensions and can be used to more accurately determine the appropriate time to perform maintenance on each offshore wind turbine.

[0137] S3: With the goal of minimizing the total cost of each maintenance task within the maintenance cycle, a dynamic maintenance large model for offshore wind turbines is established, and a dynamic maintenance small model for each maintenance vessel is obtained based on the dynamic maintenance large model.

[0138] It should be noted that the dynamic maintenance model is a mathematical model aimed at minimizing the total cost of each maintenance task within the maintenance cycle (including labor costs of maintenance personnel, rental costs of maintenance vessels, equipment replacement and repair costs, etc.). This model considers various factors such as the operating status of the wind turbine, maintenance time, and maintenance resources (such as maintenance vessels and spare parts). Through comprehensive analysis and optimization of these factors, the optimal maintenance strategy is determined. The maintenance cycle can be preset as a fixed scheduling cycle, such as 7 days or 30 days, based on the offshore wind farm's operation and maintenance management strategy; alternatively, it can be dynamically adjusted based on the wind turbine fault prediction cycle and marine weather forecast cycle.

[0139] The dynamic maintenance mini-model is based on the dynamic maintenance macro-model. It further decomposes the maintenance task optimization model for each maintenance vessel according to the specific situation of each maintenance vessel and the assigned maintenance tasks.

[0140] S4: For the units to be maintained that fall within the corresponding multi-level maintenance time window on the current day, the units to be maintained are divided into zones based on the number of maintenance vessels dispatched on that day, and a maintenance task for each zone is assigned to each maintenance vessel.

[0141] It should be noted that the units awaiting maintenance refer to offshore wind turbines that fall within their multi-level maintenance time window on that day and require maintenance operations. Based on the actual number of maintenance vessels dispatched that day, the units awaiting maintenance are divided into several zones, with each zone's units being maintained by a single maintenance vessel. The purpose of this zoning is to rationally allocate maintenance resources, improve maintenance efficiency, and ensure that each maintenance vessel can complete its maintenance tasks within its operational capabilities.

[0142] S5: Solve the dynamic maintenance mini-model corresponding to each partition maintenance task to obtain the operation and maintenance strategy of the unit to be maintained.

[0143] It should be noted that the operation and maintenance strategy refers to the specific maintenance plan obtained by solving the dynamic maintenance mini-model for each partition's units to be maintained.

[0144] This embodiment provides a wind turbine cluster operation and maintenance method that considers multiple time windows and intelligent zoning. This method comprehensively considers various operation and maintenance needs of offshore wind turbines, including equipment status, marine environment, and economic benefits. By constructing multi-level operation and maintenance time windows and a dynamic maintenance model, it achieves refined management and optimization of wind turbine cluster operation and maintenance tasks. In specific implementation, this method first determines the opportunity maintenance interval for each offshore wind turbine. This interval is derived based on real-time equipment status monitoring data, historical maintenance records, and operating environment factors, ensuring the rationality and effectiveness of maintenance time. Subsequently, considering the safety and economic requirements of maintenance tasks, multi-level operation and maintenance time windows for each turbine are further determined, providing a time reference for subsequent operation and maintenance decisions.

[0145] When constructing the dynamic maintenance model, this method aims to minimize the total cost of each maintenance task within the maintenance cycle. It comprehensively considers various factors such as the operating status of the wind turbine, maintenance time, and maintenance resources. Through comprehensive analysis and optimization of these factors, the optimal maintenance strategy is derived. This strategy not only effectively reduces maintenance costs but also improves the reliability and power generation efficiency of the wind turbine, providing a strong guarantee for the long-term stable operation of the wind farm.

[0146] Furthermore, this method employs an intelligent zoning strategy, dividing the units to be maintained into several zones based on the actual number of maintenance vessels dispatched that day, and assigning a maintenance task to each zone. This zoning approach not only achieves a rational allocation of maintenance resources but also improves maintenance efficiency, ensuring that each maintenance vessel can complete its maintenance tasks within its operational capabilities.

[0147] In summary, the wind turbine cluster operation and maintenance method considering multiple time windows and intelligent partitioning provided in this embodiment achieves refined management and optimization of wind turbine cluster operation and maintenance tasks by comprehensively considering various operation and maintenance needs and factors, providing strong support for the long-term stable operation of wind farms.

[0148] like Figure 2 As shown, the preventative opportunistic maintenance strategy refers to a strategy that, when a certain offshore wind turbine reaches a reliability threshold and requires maintenance or repair, provides another wind turbine within the opportunistic maintenance interval with a maintenance opportunity to reduce the number of outbound maintenance trips and maintenance costs. This can significantly reduce operation and maintenance costs. In the preventative opportunistic maintenance strategy, the opportunistic maintenance interval is dynamically determined based on the actual operating status of the wind turbines and historical maintenance data. When a turbine reaches a reliability threshold and requires maintenance or repair, the system assesses the operating status and location information of other turbines, selects those within the opportunistic maintenance interval, and assigns maintenance tasks to them. This strategy not only reduces the number of outbound maintenance vessels and lowers operation and maintenance costs but also avoids potential failures through timely maintenance, improving the overall reliability of the wind turbines. Furthermore, the preventative opportunistic maintenance strategy also considers the overall operation and maintenance needs of the wind farm, achieving the rational allocation and efficient utilization of operation and maintenance resources by optimizing the arrangement of maintenance tasks.

[0149] In one embodiment of the present invention, step S1 provides a method for determining the opportunity maintenance interval. This method obtains the opportunity maintenance interval of an offshore wind turbine by utilizing a three-parameter Weibull distribution and reliability analysis, and includes the following steps:

[0150] S11: Based on the three-parameter Weibull distribution and reliability analysis, historical failure data of each offshore wind turbine are fitted to obtain the failure probability density function of each turbine component. Failure probability function Reliability function and failure rate function The following are the details:

[0151] ;

[0152] ;

[0153] ;

[0154] ;

[0155] In the formula, t represents the cumulative operating time of the fan components. For scale parameters, For shape parameters, For position parameters.

[0156] Among them, the three-parameter Weibull distribution is a probability distribution model used to describe the distribution of product lifespan, including scale, shape, and location parameters; reliability analysis refers to the method of evaluating the probability of a product completing its specified function under specified conditions and within a specified time through statistical analysis; historical failure data refers to the recorded data such as the time, cause, and consequences of equipment failures during past operation; the failure probability density function is a function describing the rate of change of the failure probability of a product at a certain moment; the failure probability function is a function describing the probability of a product failing within a specified time; the reliability function is a function describing the probability of a product maintaining normal operating condition within a specified time; and the failure rate function is a function describing the probability of a product failing per unit time at a certain moment.

[0157] S12: Set the opportunity maintenance reliability threshold R o and preventive maintenance reliability threshold R p Substituting these values ​​into the reliability function, and simultaneously solving the failure probability density function, failure probability function, reliability function, and failure rate function, we obtain the opportunity maintenance interval [T] corresponding to the opportunity maintenance reliability threshold and the preventive maintenance reliability threshold. o T p ], T o To maintain the initial time of the opportunity, T p To maintain the opportunity until the end of the time.

[0158] Among them, the opportunity maintenance reliability threshold refers to the critical value of equipment reliability that triggers opportunity maintenance; the preventive maintenance reliability threshold refers to the critical value of equipment reliability that triggers preventive maintenance.

[0159] like Figure 3 As shown, Figure 3 This illustrates the relationship between the reliability function of offshore wind turbine components and the opportunity maintenance interval. This is achieved by setting an opportunity maintenance reliability threshold R. o and preventive maintenance reliability threshold R p This allows us to determine the corresponding opportunity maintenance interval [T]. o T p ].

[0160] Taking the classic experimental data of failures of six major components of offshore wind turbines as an example, a three-parameter Weibull distribution was fitted to calculate the reliability function parameters of the wind turbine components. At the same time, the opportunity maintenance reliability threshold and preventive maintenance reliability threshold of each type of component were set. The results are shown in Table 1.

[0161] Table 1. Three parameters of wind turbine components: Weibull parameters and reliability thresholds.

[0162]

[0163] In a further embodiment, for step S2, a method for determining a multi-level operation and maintenance time window is provided. This method includes the following steps:

[0164] S21: Determine the time window that considers the security of the maintenance task by using the security of the maintenance task as a constraint, and denote it as the first-level time window T. d .

[0165] It should be noted that the Level 1 time window refers to the highest priority maintenance time window determined by the constraint of maintenance task safety. To ensure the safety of the maintenance task, the sea conditions and weather must be safe on the day of departure; otherwise, departure will not be permitted that day. The time range within which departure cannot be made due to marine environmental factors is T. d .

[0166] S22: Define the opportunity maintenance interval as a secondary time window [T] o T p ].

[0167] It should be noted that the secondary time window refers to the second-priority maintenance time window determined by the equipment reliability status as a constraint. The opportunistic maintenance interval is a time window based on component reliability.

[0168] S23: Determine the time window for considering the economics of offshore wind power by using power generation loss as a constraint, and denot it as T. loss .

[0169] It should be noted that power generation loss refers to the economic loss caused by the reduction in power generation due to wind turbine shutdown for maintenance. To further improve the economics of offshore wind power, power generation loss should be minimized as much as possible within a maintenance cycle. This is achieved by predicting offshore wind power output and avoiding maintenance during periods of high wind power output. Let T be the time range of peak offshore wind power output. loss .

[0170] S24: Let T be the maintenance time corresponding to the initial reliability of the offshore wind turbine. c If the maintenance time T c If it is within the second-level time window, then the multi-level operation and maintenance time window is... If maintenance time T c If it is not within the second-level time window, then the multi-level operation and maintenance time window is... .

[0171] It should be noted that initial reliability refers to the reliability value of the offshore wind turbine at the current moment; maintenance time refers to the planned time point for maintenance operations on the offshore wind turbine.

[0172] This embodiment is based on three time windows [T] o T p ]、Td and T loss Determine multi-level operation and maintenance time windows. Among them, T d It is constrained by the security of maintenance tasks and is a condition that must be met in maintenance strategy modeling; this is called the first-level time window. o T p The second time window is defined by the reliability status of the wind turbine components. Failure to meet these conditions will result in under-maintenance or over-maintenance, leading to severe economic losses or even equipment damage. loss It is constrained by the loss of power generation, which affects the economics of offshore wind power to a certain extent, and is called the third-level time window. Figure 3 The horizontal axis illustrates the construction of multi-level time windows for wind turbine components. The red portion represents the first-level time window, the yellow portion the second-level time window, and the green portion the third-level time window. During maintenance tasks, the first-level time window involves the safety of maintenance personnel and must be met; otherwise, shipboard maintenance will not proceed. The second-level time window involves significant economic losses or equipment health conditions and should be met first, provided there are no safety hazards. The third-level time window should be prioritized if all other time windows are met and sufficient maintenance equipment and personnel are available, in order to improve the economic efficiency of the maintenance strategy.

[0173] In summary, let T be the maintenance time corresponding to the initial reliability of a certain offshore wind turbine. c The method for determining its multi-level operation and maintenance time window is as follows:

[0174] 1) If the fan's T c It is already in the second-level time window [T] o T p Within this timeframe, the multi-level operation and maintenance time window for the wind turbine can be determined by... It means that, among them Indicates a non-set;

[0175] 2) If the fan's T c Not yet in the second time window [T] o T p Within this timeframe, the multi-level operation and maintenance time window for the wind turbine can be determined by... It means that, among them It represents a non-set.

[0176] The dynamic update process of the multi-level operation and maintenance time window for wind turbines is as shown in the example. Figure 4 As shown. Considering the uncertainties of the marine environment and the impact of sudden wind turbine failures, the multi-level operation and maintenance time window proposed in this invention adopts a dynamic updating method. The multi-level operation and maintenance time window of the wind turbine is updated daily based on factors such as the unit's maintenance status, weather, and sea conditions.

[0177] In one embodiment of the present invention, a method for constructing a dynamic maintenance model is provided for step S3. The dynamic maintenance model is established in the following form: the total maintenance cost within a maintenance cycle is mainly composed of navigation costs. Penalty costs Maintenance costs Ship leasing costs and power generation loss The total cost of the maintenance cycle consists of five parts. As shown in the following formula:

[0178] ;

[0179] If the total cost of the entire maintenance cycle is considered The minimum value is the objective function, which makes it difficult to account for changes in the maintenance environment and the impact of sudden unit failures on maintenance strategies, and the model is complex and difficult to solve. Therefore, this invention uses multi-level operation and maintenance time windows and sudden failure information to obtain the units to be maintained on the same day, and solves the daily maintenance strategy independently based on intelligent partitioning, which can increase the flexibility of the maintenance strategy and reduce the complexity of the model.

[0180] Due to maintenance costs within a maintenance cycle It only relates to the amount of maintenance work, not the maintenance strategy. Furthermore, the maintenance task for one partition is handled by one ship; therefore, the objective function for the daily maintenance strategy of each partition is as follows:

[0181] ;

[0182] In the formula, , , and Let represent the total cost, navigation cost, penalty cost, and power generation loss of the i-th partition on that day, respectively.

[0183] 1) Flight costs :

[0184] The cost of vessel navigation depends on the type of maintenance vessel and the distance traveled. This invention assumes that all maintenance tasks use the same type of maintenance vessel, and the unit navigation cost is... The cost of navigation is as follows:

[0185] ;

[0186] In the formula, This represents the distance from wind turbine i to wind turbine j (i, j = 0, 1, ..., n), with the number 0 representing the port location; For 0-1 decision variables, as shown in the following formula:

[0187] ;

[0188] 2) Penalty Costs :

[0189] Penalty costs refer to the penalties imposed for over-maintenance or under-maintenance of the unit. The cost of maintenance is penalized per unit of time. The penalty cost per unit time for under-maintenance is shown in the following formula:

[0190] ;

[0191] ;

[0192] In the formula, This indicates the maintenance time of the i-th generating unit. This represents the upper limit of the time window for the i-th generator unit. This represents the lower limit of the time window for the i-th generator unit.

[0193] 3) Power generation loss :

[0194] Power generation loss is caused by unit downtime due to maintenance, and is related to power generation capacity and downtime. The cost per kilowatt-hour is... The power generation loss is shown in the following formula:

[0195] ;

[0196] In the formula, This represents the downtime of the i-th generating unit. This represents the generating capacity of the i-th generating unit.

[0197] The constraints of the above maintenance model are as follows:

[0198] 1) Reliability constraints:

[0199] The constraint relationship between the preventive maintenance reliability threshold and the opportunistic maintenance reliability threshold in the preventive opportunistic maintenance strategy is shown in the following formula:

[0200] ;

[0201] In the formula, This represents the preventive maintenance reliability threshold for the i-th unit. This represents the opportunity maintenance reliability threshold for the i-th unit.

[0202] 2) Daily working hours constraints:

[0203] To ensure the safety of maintenance tasks, the daily working hours are set to no more than 12 hours. The daily working time constraint is as follows:

[0204] ;

[0205] In the formula, This represents the maintenance time for the i-th generating unit. This indicates the total time the ship traveled.

[0206] In one embodiment of the present invention, for step S4, a method for intelligent partitioning is provided, comprising the following steps:

[0207] S41: Based on the multi-level maintenance time window, the units to be maintained are divided into alarm maintenance units and opportunity maintenance units; alarm maintenance units are units whose current date is the last day of the corresponding multi-level maintenance time window; opportunity maintenance units are all units in the units to be maintained except alarm maintenance units.

[0208] The three types of units involved in this step are as follows:

[0209] Units awaiting maintenance: All units that fall within their multi-level maintenance time window on the current date;

[0210] Alarm maintenance units: Units whose current date is the last day of their multi-level maintenance time window;

[0211] Opportunity maintenance units: Units awaiting maintenance, excluding alarm maintenance units.

[0212] S42: Based on the number of maintenance vessels dispatched on the day, a clustering algorithm is used to cluster the maintenance units to be maintained, and each resulting cluster is used as a partition. If there are no alarm maintenance units on the day, the opportunity maintenance units are directly clustered according to the number of vessels dispatched. If there are alarm maintenance units on the day, each alarm maintenance unit is used as a cluster center and then clustered according to the number of vessels dispatched.

[0213] It should be noted that clustering algorithm refers to an unsupervised learning algorithm that divides similar data points in a dataset into different clusters; a cluster refers to a set of similar data points obtained by the clustering algorithm; and the cluster center refers to the core data point used to represent a cluster during the clustering process.

[0214] In a further embodiment of the present invention, regarding step S42, a method for intelligent partitioning based on an intelligent partitioning maintenance strategy is provided. This method performs cluster analysis on the units to be maintained that day, centered on the alarm maintenance units, such as... Figure 5 As shown, it includes:

[0215] 1) No alarm units were detected that day:

[0216] Step 1-1. Determine the number of ships to be dispatched, n;

[0217] Step 1-2. Using the number of ships n as the cluster number, clustering is performed using clustering algorithms such as kmeans++.

[0218] Step 1-3. Assume that each ship is assigned a maintenance task for a partition, and solve for the maintenance strategy of each partition.

[0219] 2) The number of alarm units on that day equals the number of ships dispatched:

[0220] Step 2-1. Determine the number of ships to be dispatched, n;

[0221] Step 2-2. Using the number of ships n as the cluster number and the alarm groups of the day as the initial cluster centers, clustering is achieved using clustering algorithms such as kmeans;

[0222] Step 2-3. Assume that each ship is assigned a maintenance task for a partition, and solve for the maintenance strategy of each partition.

[0223] 3) The number of alarm crews on that day does not equal the number of ships dispatched:

[0224] Step 3-1. Determine the number of ships to be dispatched, n;

[0225] Step 3-2. Using the number of alarm units k as the cluster number and the alarm units of the day as the initial cluster centers, the initial clustering results are obtained using clustering algorithms such as kmeans.

[0226] Step 3-3. The clustering and aggregation process is as follows: Figure 6 As shown, based on the principle of maximizing the inter-class distance, k clusters are aggregated or decomposed into n classes;

[0227] Step 3-4. Assume that each ship is assigned a maintenance task for a partition, and solve for the maintenance strategy of each partition.

[0228] In one embodiment of the present invention, for step S5, a method for solving the model based on a genetic algorithm is provided, comprising the following steps:

[0229] S51: Initialize the 0-1 population; each particle in the 0-1 population represents an operation and maintenance scheme. Each operation and maintenance scheme satisfies the daily working time constraint and includes a decision variable indicating whether to maintain the units to be maintained in its respective partition. The decision variable is 0 to indicate no maintenance and 1 to indicate maintenance.

[0230] It should be noted that a 0-1 population refers to a population composed of multiple individuals encoded with 0-1, each individual representing a possible solution; a particle refers to a single individual in the population, corresponding to an operation and maintenance plan; an operation and maintenance plan refers to the specific maintenance arrangements formulated for the units to be maintained within a partition; daily working time constraints refer to the restrictions on the daily working time of the operation and maintenance vessels; and decision variables refer to 0-1 variables used to represent whether to perform maintenance on a certain unit.

[0231] First, assume that a certain partition has N For a given unit to be repaired, the optimization parameters are set as N 0-1 decision variables. As shown in the following formula:

[0232] ;

[0233] Replacing the initial population with a randomly generated 0-1 population that satisfies the daily working time constraint can effectively reduce the solution time. The initial population is shown in the following formula:

[0234] ;

[0235] In the formula, pop represents the initial population, and popsize represents the population size. This indicates a daily working time constraint. `random()` is a 0-1 random generation function used to randomly generate decision variables with values ​​of 0 or 1.

[0236] S52: Use partial matching crossover and adaptive mutation methods to perform crossover and mutation operations on the 0-1 population.

[0237] Among them, the partial matching crossover method refers to a crossover operation method specifically used for permutation encoding, which generates legitimate offspring by exchanging partial segments of the parent generation and repairing conflicts; the adaptive mutation method refers to a mutation operation method that dynamically adjusts the mutation probability according to the fitness of the individual; crossover and mutation operations are two basic operations in genetic algorithms used to generate new individuals. The crossover operation generates offspring by combining parent genes, and the mutation operation increases population diversity by randomly changing individual genes.

[0238] This step replaces traditional crossover with PMX, a crossover operation specifically designed for handling permutations and combinations. PMX ensures that each gene will not be repeated in the offspring and is widely used in solving the TSP problem. The PMX strategy is as follows:

[0239] Step 1. Randomly select two crossing points in a pair of parent chromosomes; the part between the two points is called the exchange segment.

[0240] Step 2. Exchange the genes of the two exchange segments;

[0241] Step 3. Perform conflict detection on the two resulting offspring.

[0242] To further improve the search efficiency of the genetic algorithm, this step also replaces the traditional mutation with adaptive mutation, which can dynamically adjust the mutation rate according to the fitness level, and can significantly improve the quality and diversity of the population.

[0243] S53: Select particles from the current population whose fitness exceeds a set value to enter the next generation of the population.

[0244] Among them, fitness value refers to a numerical value used to measure the quality of an individual, which is usually calculated based on the objective function; selection operation refers to the operation of selecting superior individuals from the current population to enter the next generation population; the next generation population refers to the population obtained after the selection operation for continued evolution.

[0245] In the offshore wind power maintenance strategy of this invention, the total operation and maintenance cost is minimized; therefore, the total cost is set as the fitness function. Considering the unit maintenance priority, it is necessary to ensure that units that alarm on the same day are maintained first. The fitness function is as follows:

[0246] ;

[0247] In the formula, This represents the maintenance scheme for the i-th partition. Let represent the set of alarm units in the i-th partition on that day. 'inf' is positive infinity and is used to mark invalid operation and maintenance plans that do not cover all alarm units. These plans will be directly eliminated during the genetic algorithm selection process.

[0248] If the partitioned maintenance scheme includes all its alarm units, then the fitness function is the total cost. Otherwise, the maintenance plan is invalid, and the fitness value is infinite.

[0249] S54: Repeatedly perform crossover and selection operations on the next generation population until the preset termination conditions are met, and determine the operation and maintenance scheme corresponding to the particle with the highest fitness in the final population as the operation and maintenance strategy of the unit to be maintained in the corresponding partition.

[0250] The preset termination condition refers to the pre-defined conditions under which the algorithm stops running, such as reaching the maximum number of iterations or the fitness value converging; the final population refers to the population obtained when the algorithm terminates.

[0251] The above method will be explained in detail below with reference to the accompanying drawings and specific embodiments.

[0252] Taking an offshore wind farm with 100 wind turbines as an example, the location layout and numbering of the wind farm are as follows: Figure 7As shown, adjacent wind turbines are spaced 1 km apart, and the port is located at the midpoint of the wind farm, 30 km away. Each wind turbine has an installed capacity of 4MW, and the wind power price is taken as 0.75 yuan / kWh, referencing China's offshore wind power grid connection price. The over-maintenance penalty cost is 3000 yuan / day, and the delayed maintenance penalty cost is 12000 yuan / day. The maximum number of vessels allowed to operate per day is limited to 2, the maintenance vessel rental fee is 20000 yuan / day, the navigation cost is 100 yuan / km, the navigation speed is 20 km / h, and the maximum daily working time is 12 hours. It is assumed that maintenance personnel meet the task requirements. The maintenance status of six key wind turbine components—pitch system, main shaft, gearbox, hydraulic system, generator, and frequency converter—and the maintenance time for each component are shown in Table 2.

[0253] Table 2 Maintenance Time for Fan Components

[0254]

[0255] To verify the economic efficiency of the maintenance strategy proposed in this invention, three scenarios were set up for comparative analysis.

[0256] (1) Scenario 1 (without considering time window): without considering time window constraints throughout the entire maintenance cycle, consider the route planning of the maintenance vessel;

[0257] (2) Scenario 2 (considering only the first and second time windows): considering the first and second time windows of the wind turbine components, that is, considering the reliability and safety constraints of the components, but not considering the power generation loss caused by wind speed differences;

[0258] (3) Scenario 3 (Strategy proposed in this invention): Consider the multi-level operation and maintenance time window of wind turbine components, and take into account the impact of component reliability, maintenance safety and wind speed.

[0259] The maintenance strategies obtained for each scenario are shown in Table 3. The total maintenance costs for scenarios 1, 2, and 3 are RMB 512,387, RMB 443,157, and RMB 421,407, respectively. It can be seen that compared to scenario 1, scenarios 2 and 3 significantly reduce maintenance costs by considering the time windows of wind turbine components, thus avoiding under-maintenance and over-maintenance. The maintenance costs for scenarios 2 and 3 are reduced by 13.51% and 17.76% compared to scenario 1, respectively. Furthermore, the solution times for scenarios 1, 2, and 3 are 6.95s, 3.47s, and 2.79s, respectively. The solution speeds for scenarios 2 and 3 are improved by 100.28% and 149.06% compared to scenario 1, respectively. This is because scenarios 2 and 3 adopt an intelligent partitioning strategy to decompose a single large model into multiple smaller models, reducing the complexity of a single model and effectively improving the overall solution speed.

[0260] Table 3 Maintenance Strategies for Each Scenario

[0261]

[0262] The maintenance strategy Gantt chart obtained in Scenario 2 is as follows: Figure 8 As shown, the maintenance strategy Gantt chart obtained in scenario 3 is as follows: Figure 9 As shown in the figure, both Scenario 2 and Scenario 3 strategies ensure that the maintenance time of the units to be maintained is within the time window. However, Scenario 2 does not include a third-level time window, and maintenance is still carried out on the second day when power generation is relatively high. Scenario 3, while satisfying the first and second-level time windows, also takes into account the impact of wind speed, avoiding maintenance on the second day and significantly reducing power generation loss. As shown in Table 3, with other costs being roughly the same, the power generation loss in Scenario 2 is 142,249 yuan, while the power generation loss in Scenario 3 is 120,986 yuan, a relative reduction of 14.95%. In summary, the maintenance strategy proposed in this invention comprehensively considers multiple conditions such as component reliability, sea conditions, and meteorological factors, making it more adaptable to the complex and ever-changing marine environment. This ensures the safety of maintenance tasks and improves the economic benefits of offshore wind power.

[0263] In summary, compared with the prior art, the present invention has at least the following advantages and beneficial effects:

[0264] (1) The maintenance of offshore wind turbines is affected by multiple factors. The multi-level operation and maintenance time window proposed in this invention takes into account the influence of component status, environment and weather, which not only ensures the safety of maintenance tasks, but also improves their economy.

[0265] (2) By using the intelligent partitioning strategy to decompose a single maintenance model into multiple smaller models, the complexity of a single model is reduced, which can effectively improve the solution speed of the overall maintenance strategy.

[0266] (3) The daily maintenance strategy can be dynamically updated based on the unit maintenance information and environmental change information, which can effectively improve the flexibility of the maintenance strategy and is more suitable for offshore wind turbines with variable environment and sudden failure.

[0267] Based on the same inventive concept, embodiments of the present invention also provide a wind turbine group operation and maintenance device that considers multiple time windows and intelligent partitioning for implementing the wind turbine group operation and maintenance method that considers multiple time windows and intelligent partitioning as described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in the embodiments of the wind turbine group operation and maintenance device considering multiple time windows and intelligent partitioning provided below can be found in the limitations of the wind turbine group operation and maintenance method considering multiple time windows and intelligent partitioning described above, and will not be repeated here.

[0268] Please see Figure 10 This invention also provides a wind turbine group operation and maintenance device that considers multiple time windows and intelligent zoning, including:

[0269] The multi-time window determination module is used to determine the opportunity maintenance interval for each offshore wind turbine; it is also used to determine the time window for each offshore wind turbine considering the safety of maintenance tasks and the time window considering the economics of offshore wind power, and to obtain the multi-level operation and maintenance time window for each offshore wind turbine by combining the opportunity maintenance interval and the time corresponding to the determined time window.

[0270] The model building module is used to establish a dynamic maintenance large model of offshore wind turbines with the goal of minimizing the total cost of each maintenance task within the maintenance cycle, and to decompose the dynamic maintenance small model of each maintenance vessel based on the dynamic maintenance large model.

[0271] The intelligent partitioning module is used to partition the units to be maintained according to the number of maintenance vessels dispatched on the same day, based on the corresponding multi-level maintenance time window, and assign a maintenance task to each maintenance vessel for that partition.

[0272] The model solving module is used to solve the dynamic maintenance mini-model corresponding to each partition maintenance task to obtain the operation and maintenance strategy of the unit to be maintained.

[0273] In one embodiment of the present invention, in the step of determining the opportunity maintenance interval of each offshore wind turbine, the opportunity maintenance interval of the offshore wind turbine is obtained by utilizing a three-parameter Weibull distribution and reliability analysis.

[0274] The opportunity maintenance interval for offshore wind turbines is obtained using a three-parameter Weibull distribution and reliability analysis, including:

[0275] Based on the three-parameter Weibull distribution and reliability analysis, the historical fault data of each offshore wind turbine were fitted to obtain the failure probability density function, failure probability function, reliability function and failure rate function of each wind turbine component;

[0276] Set the opportunity maintenance reliability threshold and the preventive maintenance reliability threshold, and substitute them into the reliability function. Solve the failure probability density function, failure probability function, reliability function and failure rate function simultaneously to obtain the opportunity maintenance interval corresponding to the opportunity maintenance reliability threshold and the preventive maintenance reliability threshold.

[0277] In one embodiment of the present invention, the step of determining a time window considering the safety of maintenance tasks and a time window considering the economics of offshore wind power for each offshore wind turbine, and combining the opportunity maintenance interval and the time corresponding to the determined time window to obtain the multi-level operation and maintenance time window for each offshore wind turbine includes:

[0278] The time window that considers the security of the maintenance task is determined by using the security of the maintenance task as a constraint, and is denoted as the first-level time window T. d ;

[0279] The opportunity maintenance interval is denoted as the secondary time window [T]. o T p ], T o To maintain the initial time of the opportunity, T p To maintain the opportunity at the end of the time;

[0280] The time window for considering the economics of offshore wind power is determined by using power generation loss as a constraint, and denoted as T. loss ;

[0281] Let T be the maintenance time corresponding to the initial reliability of the offshore wind turbine. c If the maintenance time T c If it is within the second-level time window, then the multi-level operation and maintenance time window is... If maintenance time T c If it is not within the second-level time window, then the multi-level operation and maintenance time window is... .

[0282] In one embodiment of the present invention, the expressions for dynamically maintaining the large model and dynamically maintaining the small model are as follows:

[0283] The objective function for dynamically maintaining the large model is as follows:

[0284] ;

[0285] In the formula, For total cost, For the cost of travel, To incur penalties, To maintain costs, For vessel leasing costs, This is due to the loss of power generation.

[0286] The objective function for dynamically maintaining the small model is as follows:

[0287] ;

[0288] In the formula, , , and Let represent the total cost, navigation cost, penalty cost, and power generation loss of the i-th partition on that day, respectively;

[0289] The constraints of the model are as follows:

[0290] Reliability constraints:

[0291] ;

[0292] In the formula, This represents the preventive maintenance reliability threshold for the i-th unit. This represents the opportunity maintenance reliability threshold for the i-th generating unit;

[0293] Daily working hours constraints:

[0294] ;

[0295] In the formula, This represents the maintenance time for the i-th generating unit. This indicates the total sailing time of the maintenance vessel, with 12 hours representing the daily working time.

[0296] In one embodiment of the present invention, the step of partitioning the units to be maintained according to the number of maintenance vessels dispatched on that day, for units that fall within the corresponding multi-level maintenance time window on a given date, includes:

[0297] Based on the multi-level maintenance time window, the units to be maintained are divided into alarm maintenance units and opportunity maintenance units; alarm maintenance units are units whose current date is the last day of the corresponding multi-level maintenance time window; opportunity maintenance units are all units in the units to be maintained except alarm maintenance units;

[0298] Based on the number of maintenance vessels dispatched on that day, a clustering algorithm is used to cluster the maintenance units to be maintained, and each resulting cluster is used as a partition. If there are no alarm maintenance units on that day, the opportunity maintenance units are directly clustered according to the number of vessels dispatched. If there are alarm maintenance units on that day, each alarm maintenance unit is used as a cluster center and then clustered according to the number of vessels dispatched.

[0299] In one embodiment of the present invention, the step of clustering the units to be maintained using a clustering algorithm based on the number of maintenance vessels dispatched on that day includes:

[0300] If there are no alarm maintenance units on that day, the number of ships dispatched will be used as the cluster number, and clustering will be achieved using a clustering algorithm.

[0301] If the number of alarm maintenance units on the same day is equal to the number of ships dispatched, then the number of ships dispatched is used as the cluster number, and the alarm maintenance units on the same day are used as the initial cluster centers. Clustering is achieved using a clustering algorithm.

[0302] If the number of alarm maintenance units on a given day is not equal to the number of ships dispatched, then the number of alarm maintenance units is used as the cluster number, and the alarm maintenance units on that day are used as the initial cluster centers. The initial clustering results are obtained using a clustering algorithm. Based on the principle of maximizing the inter-cluster distance, several clusters in the initial clustering results are aggregated or decomposed according to the number of ships dispatched.

[0303] In one embodiment of the present invention, in the step of solving the dynamic maintenance mini-model corresponding to each partition maintenance task to obtain the operation and maintenance strategy of the unit to be maintained, an improved genetic algorithm is used for solving, and the solution process includes the following steps:

[0304] Initialize a 0-1 population; each particle in the 0-1 population represents an operation and maintenance scheme. Each operation and maintenance scheme satisfies the daily working time constraint and includes a decision variable indicating whether to maintain the units to be maintained in its respective partition. Here, a decision variable of 0 indicates no maintenance and a decision variable of 1 indicates maintenance.

[0305] A partial matching crossover method and an adaptive mutation method were used to perform crossover and mutation operations on the 0-1 population.

[0306] The selection operation is based on the fitness value, selecting particles with a fitness value exceeding the set value from the current population to enter the next generation of the population;

[0307] Repeatedly perform crossover and selection operations on the next generation population until the preset termination conditions are met, and determine the operation and maintenance scheme corresponding to the particle with the highest fitness in the final population as the operation and maintenance strategy of the unit to be maintained in the corresponding partition.

[0308] Those skilled in the art will understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the functions described above can be assigned to different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this invention. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0309] Reference Figure 11 This invention also provides a computer device, including: a memory and a processor, and a computer program stored in the memory. When the computer program is executed on the processor, it performs the following steps:

[0310] Determine the opportunity maintenance interval for each offshore wind turbine;

[0311] Determine the time window for each offshore wind turbine to consider the safety of maintenance tasks and the time window to consider the economics of offshore wind power. Combine the opportunity maintenance interval and the time corresponding to the determined time window to obtain the multi-level operation and maintenance time window for each offshore wind turbine.

[0312] With the goal of minimizing the total cost of each maintenance task within the maintenance cycle, a dynamic maintenance large model for offshore wind turbines is established, and a dynamic maintenance small model for each maintenance vessel is obtained based on the dynamic maintenance large model.

[0313] For the units to be maintained that fall within the corresponding multi-level maintenance time window on that day, the units to be maintained are divided into zones based on the number of maintenance vessels dispatched that day, and a maintenance task for each zone is assigned to each maintenance vessel.

[0314] The dynamic maintenance mini-model corresponding to each partition maintenance task is solved to obtain the operation and maintenance strategy of the unit to be maintained.

[0315] The computer device may be a desktop computer, laptop, handheld computer, or cloud server, etc. This computer device may include, but is not limited to, a processor and memory. Those skilled in the art will understand that... Figure 11 The examples of computer devices are merely examples and do not constitute a limitation on computer devices. They may include more or fewer components than shown in the illustration, or combinations of certain components, or different components. For example, they may also include input / output devices, network access devices, etc.

[0316] The processor referred to can be a Central Processing Unit (CPU), but it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor.

[0317] In some embodiments, the memory may be an internal storage unit of the computer device, such as a hard drive or RAM. In other embodiments, the memory may be an external storage device of the computer device, such as a plug-in hard drive, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card. Furthermore, the memory may include both internal and external storage units of the computer device. The memory is used to store the operating system, applications, boot loader, data, and other programs, such as the program code of the computer program. The memory can also be used to temporarily store data that has been output or will be output.

[0318] This invention also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, performs the following steps:

[0319] Determine the opportunity maintenance interval for each offshore wind turbine;

[0320] Determine the time window for each offshore wind turbine to consider the safety of maintenance tasks and the time window to consider the economics of offshore wind power. Combine the opportunity maintenance interval and the time corresponding to the determined time window to obtain the multi-level operation and maintenance time window for each offshore wind turbine.

[0321] With the goal of minimizing the total cost of each maintenance task within the maintenance cycle, a dynamic maintenance large model for offshore wind turbines is established, and a dynamic maintenance small model for each maintenance vessel is obtained based on the dynamic maintenance large model.

[0322] For the units to be maintained that fall within the corresponding multi-level maintenance time window on that day, the units to be maintained are divided into zones based on the number of maintenance vessels dispatched that day, and a maintenance task for each zone is assigned to each maintenance vessel.

[0323] The dynamic maintenance mini-model corresponding to each partition maintenance task is solved to obtain the operation and maintenance strategy of the unit to be maintained.

[0324] In this embodiment, if the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include at least: any entity or device capable of carrying computer program code to a photographic device / terminal device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electrical carrier signals or telecommunication signals.

[0325] This invention provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0326] Determine the opportunity maintenance interval for each offshore wind turbine;

[0327] Determine the time window for each offshore wind turbine to consider the safety of maintenance tasks and the time window to consider the economics of offshore wind power. Combine the opportunity maintenance interval and the time corresponding to the determined time window to obtain the multi-level operation and maintenance time window for each offshore wind turbine.

[0328] With the goal of minimizing the total cost of each maintenance task within the maintenance cycle, a dynamic maintenance large model for offshore wind turbines is established, and a dynamic maintenance small model for each maintenance vessel is obtained based on the dynamic maintenance large model.

[0329] For the units to be maintained that fall within the corresponding multi-level maintenance time window on that day, the units to be maintained are divided into zones based on the number of maintenance vessels dispatched that day, and a maintenance task for each zone is assigned to each maintenance vessel.

[0330] The dynamic maintenance mini-model corresponding to each partition maintenance task is solved to obtain the operation and maintenance strategy of the unit to be maintained.

[0331] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0332] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this invention can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0333] In the embodiments disclosed in this invention, it should be understood that the disclosed devices / terminal equipment and methods can be implemented in other ways. For example, the device / terminal equipment embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0334] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A wind turbine cluster operation and maintenance method considering multiple time windows and intelligent zoning, characterized in that, Includes the following steps: Determine the opportunity maintenance interval for each offshore wind turbine; Determine the time window for each offshore wind turbine considering the safety of maintenance tasks and the time window considering the economics of offshore wind power, and combine the opportunity maintenance interval and the time corresponding to the determined time window to obtain the multi-level operation and maintenance time window for each offshore wind turbine. With the goal of minimizing the total cost of each maintenance task within the maintenance cycle, a dynamic maintenance large model for offshore wind turbines is established, and a dynamic maintenance small model for each maintenance vessel is obtained based on the dynamic maintenance large model. For the maintenance units that fall within the corresponding multi-level maintenance time window on the current day, the maintenance units are divided into zones based on the number of maintenance vessels dispatched on that day, and a maintenance task for each zone is assigned to each maintenance vessel. The operation and maintenance strategy of the unit to be maintained is obtained by solving the dynamic maintenance mini-model corresponding to each partition maintenance task.

2. The wind turbine group operation and maintenance method considering multiple time windows and intelligent partitioning according to claim 1, characterized in that, In the step of determining the opportunity maintenance interval for each offshore wind turbine, the opportunity maintenance interval for the offshore wind turbine is obtained by utilizing the three-parameter Weibull distribution and reliability analysis; The opportunity maintenance interval for offshore wind turbines is obtained using a three-parameter Weibull distribution and reliability analysis, including: Based on the three-parameter Weibull distribution and reliability analysis, the historical fault data of each offshore wind turbine were fitted to obtain the failure probability density function, failure probability function, reliability function and failure rate function of each wind turbine component; Set the opportunity maintenance reliability threshold and the preventive maintenance reliability threshold, and substitute them into the reliability function. Solve the failure probability density function, the failure probability function, the reliability function and the failure rate function simultaneously to obtain the opportunity maintenance interval corresponding to the opportunity maintenance reliability threshold and the preventive maintenance reliability threshold.

3. The wind turbine group operation and maintenance method considering multiple time windows and intelligent partitioning according to claim 1, characterized in that, The steps of determining the time window for each offshore wind turbine considering the safety of maintenance tasks and the time window considering the economics of offshore wind power, and combining the opportunity maintenance interval with the time corresponding to the determined time window to obtain the multi-level operation and maintenance time window for each offshore wind turbine, include: The time window that considers the security of the maintenance task is determined by constraining the security of the maintenance task, and is denoted as the first-level time window. T d ; The opportunity maintenance interval is denoted as the secondary time window. T o , T p ], T o To maintain the starting moment for the opportunity, T p To maintain the opportunity at the end of the time; The time window considering the economics of offshore wind power is determined by using power generation loss as a constraint, and is denoted as... T loss ; Let the maintenance time corresponding to the initial reliability of the offshore wind turbine be denoted as . T c If the maintenance time T c If it falls within the secondary time window, then the multi-level operation and maintenance time window is... If the maintenance time T c If it is not within the secondary time window, then the multi-level operation and maintenance time window is... .

4. The wind turbine group operation and maintenance method considering multiple time windows and intelligent partitioning according to claim 1, characterized in that, The expressions for the dynamically maintained large model and the dynamically maintained small model are as follows: The objective function for dynamically maintaining the large model is as follows: ； In the formula, For total cost, For the cost of travel, To incur penalties, To maintain costs, For vessel leasing costs, This is due to the loss of power generation. The objective function for dynamically maintaining the small model is as follows: ； In the formula, , , and They represent the number of the day. i Total cost, navigation cost, penalty cost, and power generation loss for each zone; The constraints of the model are as follows: Reliability constraints: ； In the formula, Indicates the first i The reliability threshold for preventive maintenance of the Taiwanese unit. Indicates the first i Opportunity maintenance reliability threshold for the unit; Daily working hours constraints: ； In the formula, Indicates the first i Maintenance time for the unit This indicates the total sailing time of the maintenance vessel, with 12 hours representing the daily working time.

5. The wind turbine group operation and maintenance method considering multiple time windows and intelligent partitioning according to claim 1, characterized in that, The step of partitioning the units to be maintained within the corresponding multi-level maintenance time window for the given date, based on the number of maintenance vessels dispatched that day, includes: The units to be maintained are divided into alarm maintenance units and opportunity maintenance units according to the multi-level operation and maintenance time window; the alarm maintenance units are the units whose current date is the last day of the corresponding multi-level operation and maintenance time window; the opportunity maintenance units are all the units to be maintained except for the alarm maintenance units. Based on the number of vessels dispatched for maintenance on that day, a clustering algorithm is used to cluster the units to be maintained, and each resulting cluster is used as a partition. If there are no alarm maintenance units on that day, the opportunity maintenance units are directly clustered according to the number of vessels dispatched. If there are alarm maintenance units on that day, each alarm maintenance unit is used as a cluster center and then clustered according to the number of vessels dispatched.

6. The wind turbine group operation and maintenance method considering multiple time windows and intelligent partitioning according to claim 5, characterized in that, The step of clustering the units to be maintained using a clustering algorithm based on the number of maintenance vessels dispatched that day includes: If there are no alarm maintenance units on that day, then the number of ships dispatched is used as the cluster number, and clustering is achieved using a clustering algorithm; If the number of alarm maintenance units on that day is equal to the number of ships dispatched, then the number of ships dispatched is used as the cluster number, and the alarm maintenance units on that day are used as the initial cluster centers. Clustering is achieved using a clustering algorithm. If the number of alarm maintenance units on a given day is not equal to the number of ships dispatched, then the number of alarm maintenance units is used as the cluster number, and the alarm maintenance units on that day are used as the initial cluster centers. The initial clustering results are obtained using a clustering algorithm. Based on the principle of maximizing the inter-cluster distance, several clusters in the initial clustering results are aggregated or decomposed according to the number of ships dispatched.

7. The wind turbine group operation and maintenance method considering multiple time windows and intelligent partitioning according to claim 1, characterized in that, In the step of solving the dynamic maintenance mini-model corresponding to each partition maintenance task to obtain the operation and maintenance strategy of the unit to be maintained, an improved genetic algorithm is used for the solution. The solution process includes the following steps: Initialize a 0-1 population; each particle in the 0-1 population represents an operation and maintenance scheme, each operation and maintenance scheme satisfies the daily working time constraint, and each includes a decision variable indicating whether the unit to be maintained in its partition is maintained, wherein the decision variable is 0 to indicate no maintenance and 1 to indicate maintenance; The 0-1 population was subjected to crossover and mutation operations using a partial matching crossover method and an adaptive mutation method. The selection operation is based on the fitness value, selecting particles with a fitness value exceeding the set value from the current population to enter the next generation of the population; The crossover and selection operations are repeatedly performed on the next generation population until the preset termination conditions are met, and the operation and maintenance scheme corresponding to the particle with the highest fitness in the final population is determined as the operation and maintenance strategy of the unit to be maintained in the partition.

8. A wind turbine group operation and maintenance device considering multiple time windows and intelligent zoning, characterized in that, include: The multi-time window determination module is used to determine the opportunity maintenance interval for each offshore wind turbine. It is also used to determine the time window for each offshore wind turbine considering the safety of maintenance tasks and the time window considering the economics of offshore wind power, and to obtain the multi-level operation and maintenance time window for each offshore wind turbine by combining the opportunity maintenance interval and the time corresponding to the determined time window. The model building module is used to establish a dynamic maintenance large model of offshore wind turbines with the goal of minimizing the total cost of each maintenance task within the maintenance cycle, and to decompose the dynamic maintenance small model of each maintenance vessel based on the dynamic maintenance large model. The intelligent partitioning module is used to partition the units to be maintained that fall within the corresponding multi-level maintenance time window on the current day, based on the number of maintenance vessels dispatched on that day, and assign a maintenance task to each maintenance vessel for that partition. The model solving module is used to solve the dynamic maintenance mini-model corresponding to each partition maintenance task to obtain the operation and maintenance strategy of the unit to be maintained.

9. A computer device, characterized in that, The device includes a processor and a memory: The memory is used to store computer programs and send the instructions of the computer programs to the processor; The processor executes, according to the instructions of the computer program, a wind turbine group operation and maintenance method considering multiple time windows and intelligent partitioning as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements a wind turbine cluster operation and maintenance method considering multiple time windows and intelligent partitioning as described in any one of claims 1-7.

Images