Operation and maintenance management method and platform of photovoltaic power generation system

Abstract

The invention provides an operation and maintenance management method and platform for a photovoltaic power generation system, and relates to the technical field of photovoltaic power station operation and maintenance, and the method comprises the steps: obtaining a to-be-processed operation and maintenance task set and an idle operation and maintenance personnel set; scheduling optimization is carried out on operation and maintenance tasks through an improved NSGA-II algorithm, a multi-objective optimization problem of total operation and maintenance cost minimization, total task completion time minimization and response time minimization is established, multi-objective optimization solution is carried out by adopting a clustering enhancement coding structure, a mixed population initialization strategy and a probability weighting genetic operator, and a Pareto optimal solution set is output; and selecting a final scheme from the Pareto optimal solution set, converting the final scheme into an executable format, and outputting a complete operation and maintenance scheduling scheme comprising a personnel scheduling list, a path planning instruction, a material and tool list and a total time schedule. According to the invention, multi-objective collaborative optimization of the operation and maintenance cost, the response time and the execution efficiency can be realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of photovoltaic power station operation and maintenance technology, and particularly relates to a photovoltaic power generation system operation and maintenance management method and platform. BACKGROUND

[0002] At present, the operation and maintenance strategy optimization of photovoltaic power generation system mainly focuses on two aspects: one is the operation and maintenance decision optimization based on data analysis, which collects power station operation data to perform power generation prediction, equipment fault analysis, etc., to provide data support for operation and maintenance decision; the other is the intelligent upgrading of operation and maintenance equipment, especially the application of unmanned operation and maintenance equipment. The existing technology has made certain progress in operation and maintenance cost control, response time optimization and execution efficiency improvement, but still has technical bottlenecks in multi-objective collaborative optimization, personalized strategy formulation and dynamic adaptability.

[0003] The patent for invention with Chinese publication number CN111046321A discloses a photovoltaic power station operation and maintenance strategy optimization method and device, which obtains operation and maintenance system strategy execution result information set, determines the optimization result information set with the optimal solution of preset dimensions such as operation and maintenance benefit, operation and maintenance precision and response efficiency as the target, calculates the strategy optimization factor based on the iteration factor, and iteratively optimizes the operation and maintenance strategy accordingly. The method adopts a double closed-loop iteration mechanism, the outer layer realizes the iterative optimization of the operation and maintenance strategy, and the inner layer realizes the self-learning iteration of the strategy optimization factor and the optimal solution algorithm set. The calculation of the strategy optimization factor combines the single power station optimization factor in the time dimension and the multi-power station optimization factor in the time-space dimension. However, the patent mainly aims at the macro optimization of the strategy level, iteratively improves based on the historical execution result, and lacks the real-time optimization capability for specific operation and maintenance task allocation and personnel scheduling. In addition, although the method considers multiple optimization dimensions, it does not establish an accurate multi-objective mathematical model, and it is difficult to handle the collaborative optimization problem of multiple conflicting objectives in operation and maintenance scheduling. SUMMARY

[0004] Therefore, the present application provides a photovoltaic power generation system operation and maintenance management method and platform, which solves the technical problems in the prior art, such as the coding structure being prone to produce invalid solutions with mismatched skills, the population initialization being completely random to cause slow convergence speed, the genetic operator being too random to lack directional search ability, and the lack of solution diversity protection mechanism being easy to fall into local optimum, by adopting a "list of lists" coding structure enhanced by a "dynamic clustering label" and a triple algorithm improved strategy, and realizes the multi-objective collaborative optimization of operation and maintenance cost, response time and execution efficiency.

[0005] The technical scheme of the present application is implemented as follows: On the one hand, the present application provides a photovoltaic power generation system operation and maintenance management method, comprising: S1, obtaining a set of operation and maintenance tasks to be processed and a collection of idle maintenance personnel Each of the operation and maintenance tasks Includes geographic location coordinates, urgency weight, estimated operation time, and resource requirements for each operations and maintenance personnel. Includes skill set, starting location coordinates, and hourly labor cost; This represents the total number of maintenance tasks. Total number of maintenance personnel; S2. The operation and maintenance tasks are scheduled and optimized by the improved NSGA-II algorithm. A multi-objective optimization problem is established to minimize the total operation and maintenance cost, the total task completion time, and the response time. The multi-objective optimization is solved by using a clustering-enhanced coding structure, a hybrid population initialization strategy, and a probability-weighted genetic operator, and the Pareto optimal solution set is output. S3. By selecting a final solution from the Pareto optimal solution set and converting it into an executable format, the output is a complete operation and maintenance scheduling solution including a personnel scheduling list, path planning instructions, a material and tool list, and an overall time plan.

[0006] Preferably, step S2 includes: S21. Set population size Maximum number of iterations Crossover probability Probability of mutation Initial temperature parameters Cluster number K and cluster preference threshold ; S22. Use the k-means algorithm to cluster the geographic coordinates of all operation and maintenance tasks to obtain K geographic clusters, and then cluster each operation and maintenance task. Assign cluster labels The clustering preference weight vector is calculated based on the Euclidean distance between the starting position of the maintenance personnel and each cluster center. A clustering-enhanced coding scheme is used to define chromosome X as... ,in To be allocated to maintenance personnel Task sequence, For this person The clustering preference weight vector is used, and clustering constraints are set to limit task allocation; S23. By establishing a multi-objective optimization function that minimizes total maintenance cost, total task completion time, and response time, the optimization objectives are determined. S24. Generate the initial population using a mixed population initialization strategy. 20% of the individuals were generated using a cluster-constrained greedy seed, and 80% of the individuals were generated using a cluster-constrained Latin hypercube sampling method. The number of iterations was set to t=0. S25, calculating the modified local fitness of each chromosome for each work order , selecting parent individuals for reproduction using a probabilistic weighted selection mechanism according to the modified local fitness and a dynamic temperature parameter ; S26, performing genetic operations to generate a child population , wherein the genetic operations include: performing a special crossover operator of task block exchange with a crossover probability ; performing a 2-opt local search mutation with a mutation probability , and applying a rollback mechanism; S27, forming a joint population by merging the parent population and the child population ; , performing crowded distance calculation and non-dominated sorting on the joint population, and selecting excellent individuals to form the next generation population ; S28, determining whether an iteration stop condition is reached, if yes, outputting the current Pareto optimal solution set, if not, updating the dynamic temperature parameter and returning to step S25 to continue iteration.

[0007] Preferably, step S22 includes: using a k-means algorithm to cluster the task locations to obtain K geographic clusters, and each task is assigned a cluster label ; wherein the cluster center is calculated as: wherein, is the cluster center of the kth geographic cluster, is the number of tasks in the kth geographic cluster, is the geographic position coordinates of the task ; calculating a preference weight vector of each maintenance personnel for each geographic cluster : wherein, is the preference weight vector of the maintenance personnel for the geographic cluster k, is the Euclidean distance function, is the starting position coordinates of the maintenance personnel , is a small constant to prevent division by zero; The final weight vector is obtained through normalization: In the formula, K is the total number of clusters. For clustering index; Set clustering constraints: Task Only when the clustering preference condition is met Only then can they be assigned to personnel. ,in This is the clustering preference threshold. For the task Clustering labels, Represents the weight vector The Each component, namely .

[0008] Preferably, the multi-objective optimization function includes: Minimize total operating costs: In the formula, Chromosomes The corresponding total operation and maintenance cost, Total number of maintenance personnel For maintenance personnel hourly labor costs To be allocated to maintenance personnel The task set, For the task The estimated working hours For unit distance transportation cost, For maintenance personnel Total path distance; Minimize total task completion time: In the formula, Chromosomes The corresponding total task completion time, This represents the total number of maintenance tasks. For the task Completion time; Minimize response time: In the formula, Chromosomes The corresponding total response time, For the task The urgency level weight.

[0009] Preferred strategies for initializing mixed populations include: The cluster constraint greedy seed generation accounts for 20% of the population, including cost priority seeds and urgency priority seeds, the cost priority seeds assign each task to the nearest maintenance personnel who meets the cluster constraint condition, and the urgency priority seeds are assigned to idle maintenance personnel who meets the cluster constraint condition in order of task urgency from high to low; The cluster constraint Latin hypercube sampling accounts for 80% of the population, taking the length of each maintenance personnel's work order as a sampling dimension, the total dimension is the number of maintenance personnel m, and Latin hypercube sampling is performed within the range of the number of task assignments to generate uniformly distributed work order length combinations, ensuring orthogonal and uniform coverage of the population in the multi-dimensional search space; The greedy task filling mechanism fills each maintenance personnel's work order with a corresponding number of specific tasks according to the work order length determined by Latin hypercube sampling, and the filling process selects tasks using a comprehensive greedy score that considers skill matching degree, cluster preference weight, and geographic distance factors, and preferentially selects tasks with the highest comprehensive score that meet the cluster constraint condition for assignment.

[0010] Preferably, in step S25, the probability weighted selection mechanism is specifically: Calculate the modified local fitness of each work order: wherein, is the modified local fitness of the maintenance personnel , is the sub-cost of the maintenance personnel , is the sub-task completion time of the maintenance personnel , is the sub-response time of the maintenance personnel , is the entropy weight coefficient, is the Shannon entropy regularization term; Calculate the selection probability by the softmax function: wherein, is the probability of being selected by the maintenance personnel , is the dynamic temperature parameter of the current iteration, is the personnel index.

[0011] Preferably, in step S26: The dedicated crossover operator uses a task block exchange method to select a continuous task segment of the maintenance personnel's work order from parent A and insert it into the maintenance personnel's work order in parent B that meets the cluster constraint; 2-opt local search mutation operator optimizes the path of the task sequence, and a rollback mechanism is adopted: when the local fitness after mutation is smaller than the local fitness before mutation, the local fitness is rolled back to the old state with a rollback probability , wherein is the local fitness after mutation, is the local fitness before mutation, is the current iteration number, is the maximum iteration number.

[0012] Preferably, the iteration stopping condition is that the maximum iteration number is reached or a convergence condition is reached, wherein the convergence condition is that convergence is determined when the change rate of the non-dominated solution set in the continuous observation window is less than a convergence threshold.

[0013] Preferably, the selection of the final execution scheme from the Pareto optimal solution set in step S3 includes two modes: an automatic selection mode automatically selects the optimal scheme from the solution set according to a preset preference; a manual selection mode presents the Pareto front in a visual manner to the operation and maintenance supervisor for manual selection.

[0014] On the other hand, the present application also provides an operation and maintenance management platform for a photovoltaic power generation system, which is used to implement the method of any one of the above aspects, and the platform comprises: a data acquisition module for acquiring photovoltaic equipment operation state data, fault alarm information, operation and maintenance personnel location information and skill resource data in real time; a task management module for receiving operation and maintenance task requirements, performing task classification, priority evaluation and resource demand analysis; an intelligent scheduling module for performing multi-objective optimization calculation based on the improved NSGA-II algorithm to generate a Pareto optimal solution set; a decision support module for providing two modes of automatic selection and manual selection to determine the final execution scheme from the Pareto optimal solution set; an execution monitoring module for outputting a complete operation and maintenance execution scheme, including a personnel scheduling list, path planning instructions, a material tool list and a time schedule, and monitoring the execution state in real time; a data storage module for storing historical operation and maintenance data, optimization parameter configuration and scheduling scheme records.

[0015] The present application has the following beneficial effects compared with the prior art: (1) The present application establishes a photovoltaic power generation system operation and maintenance management method based on the improved NSGA-II algorithm, and realizes multi-objective collaborative optimization of operation and maintenance cost, response time and execution efficiency; (2) Adopting the "dynamic clustering label" enhanced "list of lists" encoding structure, by introducing the task label based on k-means geographical clustering and the operation and maintenance personnel clustering preference weight vector in the encoding, the feasible region of task allocation is mathematically constrained; (3) By combining the hybrid initialization strategy of clustering constraint greedy seed generation and clustering constraint Latin hypercube sampling, the slow convergence problem caused by the completely random initialization of the standard NSGA-II algorithm is solved; (4) A special genetic operator suitable for the clustering enhanced encoding structure is designed, by introducing the Shannon entropy regularization term in the local fitness, the diversified allocation of tasks among different geographical clusters is encouraged, the over-concentration of tasks on a few operation and maintenance personnel is prevented, and at the same time, the rollback mechanism avoids the degradation of solution quality in the mutation process, and improves the overall quality of the Pareto front. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0017] Figure 1 Flowchart of the method of the present application Figure 2 Flowchart of the improved NSGA-II algorithm of the present application Figure 3 Schematic diagram of the platform of the present application DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0019] As shown in Figure 1 , the present application provides an operation and maintenance management method of a photovoltaic power generation system, comprising: S1, obtaining a set of operation and maintenance tasks to be processed , and a set of idle operation and maintenance personnel , wherein each operation and maintenance task contains geographical position coordinates, emergency degree weight, estimated work hours and resource demand, and each operation and maintenance personnel a set of skills, a starting position coordinate, and an hourly labor cost; total number of operation and maintenance tasks, total number of operation and maintenance personnel; S2, scheduling optimization of operation and maintenance tasks is performed by using the improved NSGA-II algorithm, a multi-objective optimization problem of minimizing total operation and maintenance cost, minimizing total task completion time, and minimizing response time is established, a clustering enhanced coding structure, a hybrid population initialization strategy, and a probability weighted genetic operator are used for multi-objective optimization solution, and a Pareto optimal solution set is output; S3, a final scheme is selected from the Pareto optimal solution set and converted into an executable format, and a complete operation and maintenance scheduling scheme including a personnel scheduling list, path planning instructions, a material and tool list, and a total time schedule is output.

[0020] Specifically, in an embodiment of the present application, step S1 includes: S11, obtaining and defining a set of operation and maintenance tasks.

[0021] First, a set of operation and maintenance tasks to be processed is obtained , wherein is the total number of operation and maintenance tasks. Each operation and maintenance task contains the following key information: geographic position coordinates: the geographic position coordinates of the task are represented by , in the form of two-dimensional coordinates , which can be GPS-based latitude and longitude coordinates or a plane coordinate system relative to a reference point.

[0022] emergency level weight: the emergency level weight of the task is represented by , which is a quantitative numerical parameter. The larger the value, the more urgent the task, which needs to be processed first. The weight value is determined according to factors such as fault type, influence range, and power generation loss.

[0023] estimated working hours: the standard operation time required to complete the task is represented by , in hours. This parameter is determined based on historical operation and maintenance data statistics and standard working hour quota of the task type.

[0024] resource requirements: the resource requirement information required by the task is represented by , including specific skill requirements, spare parts models, special tools, etc. This information is used to ensure that the assigned operation and maintenance personnel have the skills and resource conditions required to perform the task.

[0025] S12, obtaining and defining a set of operation and maintenance personnel.

[0026] Simultaneously acquire a set of available operations and maintenance personnel. ,in This represents the total number of operations and maintenance personnel. Each operations and maintenance personnel... It includes the following attribute information, among which A unique identifier for operations and maintenance personnel: Skill set: using Indicates maintenance personnel A set of skills used to meet the resource requirements of a task. Matching is performed. The skills set includes professional skills identifiers such as electrical repair, mechanical maintenance, cleaning operations, and equipment inspection.

[0027] Starting position coordinates: using Indicates maintenance personnel The starting coordinates are set using the same coordinate system as the task location coordinates. This location represents the current location of the operations and maintenance personnel or the designated starting point, and is used to calculate the travel distance and time to each task location.

[0028] Hourly labor cost: Indicates maintenance personnel The hourly labor cost includes the time-allocated value of labor cost elements such as basic wages and welfare expenses.

[0029] S13. Data integrity verification and preprocessing.

[0030] After obtaining the task set and personnel set, perform data integrity verification. For geographic coordinates, verify the validity of the coordinate values ​​and standardize the coordinate system; for urgency weights, normalize the weight values ​​to ensure they are within a reasonable range; for estimated work hours, verify their reasonableness based on task type and historical data.

[0031] In addition, a skills matching matrix was established to compare the skill sets of each operations and maintenance personnel. Resource requirements for each task Feasible personnel-task allocation combinations are pre-selected to provide constraints for subsequent optimization algorithms and avoid generating invalid solutions with skill mismatches.

[0032] Specifically, such as Figure 2 As shown, in one embodiment of the present invention, step S2 includes: S21. Set population size Maximum number of iterations Crossover probability Probability of mutation Initial temperature parameters Cluster number K and clustering preference threshold .

[0033] Specifically, in this embodiment, Set to an even number between 100 and 200, the specific value of which can be determined by the product of the task size and the number of maintenance personnel, for example... , where m is the total number of maintenance personnel and n is the total number of tasks. Set it to between 200 and 500 as one of the termination conditions of the algorithm. The value is set between 0.8 and 0.9 to control the probability of crossover operations among individuals in the population. Because this invention uses a dedicated task block exchange crossover operator, a relatively high crossover probability is beneficial for promoting the propagation of superior gene fragments while maintaining the integrity of the coding structure. The value is set between 0.1 and 0.2 to control the probability of an individual undergoing mutation. The mutation probability setting needs to consider that the 2-opt local search mutation operator used in this invention has strong local optimization capabilities; therefore, a moderate mutation probability can maintain population diversity while avoiding excessive perturbation. Set to 10.0 for the softmax function in the probability-weighted selection mechanism. A higher initial temperature helps maintain greater selection randomness in the early stages of the algorithm, preventing premature convergence. The temperature parameter will be dynamically adjusted based on the chaotic mapping during iteration. K is determined based on the geographical distribution of the tasks and the number of maintenance personnel, and is set to... The optimal number of clusters can be determined using the elbow method. Setting it to 0.3 controls the strength of geographical constraints for task assignment.

[0034] S22. Use the k-means algorithm to cluster the geographic coordinates of all operation and maintenance tasks to obtain K geographic clusters, and then cluster each operation and maintenance task. Assign cluster labels The clustering preference weight vector is calculated based on the Euclidean distance between the starting position of the maintenance personnel and each cluster center. A clustering-enhanced coding scheme is used to define chromosome X as... ,in To be allocated to maintenance personnel Task sequence, For this person The clustering preference weight vector is determined, and clustering constraints are set to limit task allocation.

[0035] Specifically, in this embodiment, step S22 includes: The k-means algorithm was used to locate the task position. Clustering is performed to obtain K geographical clusters, and each task... Assign a cluster label The specific implementation process of k-means clustering includes: firstly, randomly initializing K cluster centers, and then iteratively performing task assignment and center updating steps until convergence. The cluster center calculation formula is: In the formula, is the cluster center of the kth geographic cluster, is the number of tasks in the kth geographic cluster, is the geographic position coordinates of the task . The cluster algorithm uses Euclidean distance as the similarity measure, and the iteration termination condition is that the cluster center changes less than a preset threshold (such as 0.01) or reaches the maximum number of iterations (such as 50 times).

[0036] is calculated for each operation and maintenance personnel the preference weight vector for each geographic cluster . The weight calculation is based on the Euclidean distance between the starting position of the operation and maintenance personnel and each cluster center: In the formula, is the preference weight vector of the operation and maintenance personnel for the geographic cluster k, is the Euclidean distance function, is the starting position coordinates of the operation and maintenance personnel , is a small constant to prevent division by zero; The maximum weight vector is obtained by normalization processing: In the formula, K is the total number of clusters, is the cluster index; the normalized weight vector has the property of probability distribution, and the higher the weight value, the stronger the preference of the operation and maintenance personnel for the corresponding geographic cluster.

[0037] The "list of lists" encoding structure enhanced by clustering is used to represent the scheduling scheme. Chromosome X is defined as: wherein is the task sequence assigned to the operation and maintenance personnel , is the cluster preference weight vector of the personnel , and m is the number of operation and maintenance personnel.

[0038] This encoding structure ensures and when . Among them, indicates that the union of the task lists assigned to all operation and maintenance personnel is equal to the entire task set. when (When) indicates that there is no overlap between the task lists of different operation and maintenance personnel, i.e., an empty set, and b is the personnel index.

[0039] Set clustering constraints: Task Only when the clustering preference condition is met Only then can they be assigned to personnel. ,in This is the clustering preference threshold. For the task Clustering labels, Represents the weight vector The Each component, namely This constraint mathematically restricts the feasible domain of task allocation, avoiding allocation schemes with overly dispersed geographical locations. It effectively narrows the solution space containing "mathematically feasible but practically ineffective" solutions, thereby improving the algorithm's search efficiency and the practicality of the solutions.

[0040] It should be noted that, It is a task The cluster label is an integer value, such as 1, 2, 3..., which represents the task. Which geographic cluster does it belong to? Where K is the total number of clusters, and k is the cluster index, which is also an integer, used to represent the k-th cluster. If the task... Clustering tags If this task belongs to the third cluster, then for The explanation is that it involves taking the weight vector. The Each component, for example ,but .

[0041] S23. By establishing a multi-objective optimization function that minimizes total maintenance cost, total task completion time, and response time, the optimization objectives are determined.

[0042] Specifically, in this embodiment, the multi-objective optimization function includes: The objective function for minimizing total maintenance costs It includes two components: labor costs and transportation costs. In the formula, Chromosomes The corresponding total operation and maintenance cost, Total number of maintenance personnel For maintenance personnel hourly labor costs To be allocated to maintenance personnel The task set, for the task estimated job hours, unit distance transportation cost, total path distance for the maintenance personnel ; The artificial cost part calculates the total artificial cost required for all maintenance personnel to complete their assigned tasks, and the transportation cost part calculates the total transportation cost during the execution of tasks. The path distance is calculated according to the positions of adjacent tasks in the task sequence and the starting position.

[0043] Total task completion time minimization objective function Also known as "makespan", it represents the total time required to complete all tasks: In the formula, denotes the chromosome corresponding total task completion time, is the total number of maintenance tasks, is the completion time of the task ; The calculation of the task completion time needs to consider the departure time of the maintenance personnel executing the task, the transportation time to the task site, and the cumulative work time and transportation time of all other tasks before the personnel executing the task .

[0044] Response time minimization objective function Consider the emergency weight of the task to ensure that high emergency tasks can be given priority response: In the formula, denotes the chromosome corresponding total response time, is the emergency weight of the task .

[0045] This objective function effectively punishes scheduling schemes that cause high-emergency tasks to be completed late by multiplying the completion time of each task by its emergency weight and summing them. The higher the emergency weight of the task, the greater the negative impact of its delayed completion on the objective function value, thereby guiding the algorithm to prioritize the timely completion of emergency tasks.

[0046] The three objective functions constitute a multi-objective optimization problem: The solution of the multi-objective optimization problem is a set of Pareto optimal solutions, each of which represents a different trade-off scheme between cost, efficiency, and responsiveness. Since the three objectives conflict with each other, there is no single solution that optimizes all objectives simultaneously, so it is necessary to find multiple non-dominated solutions on the Pareto front through the NSGA-II algorithm.

[0047] S24, generating an initial population by a mixed population initialization strategy , 20% of the individuals are generated by cluster-constrained greedy seed generation, and 80% of the individuals are generated by cluster-constrained Latin hypercube sampling, and the number of iterations t is set to 0.

[0048] Specifically, in this embodiment, the mixed population initialization strategy includes: Cluster-constrained greedy seed generation accounts for 20% of the population, including cost-priority seed and urgency-priority seed, each accounting for 10%.

[0049] The cost-priority seed assigns each task to the nearest maintenance personnel who meets the cluster constraint condition, and the specific process is as follows: 1. Create a candidate task list, initialized as the set of all tasks . 2. For each unassigned task , calculate the assignment cost between it and all maintenance personnel: , where is the current position of the maintenance personnel (initially , updated to the position of the last assigned task after assignment). 3. Under the premise of meeting the cluster constraint condition and the skill matching condition , select the personnel-task pair with the minimum assignment cost for assignment. 4. Update the current position of the maintenance personnel and the candidate task list, and repeat the above process until all tasks are assigned.

[0050] The urgency-priority seed assigns tasks to idle maintenance personnel in order of task urgency from high to low that meet the cluster constraint condition; the specific process is as follows: 1. Sort all tasks in order of task urgency weight from high to low. 2. Select an idle maintenance personnel that meets the cluster constraint condition and skill matching for each task in turn. 3. If there are multiple maintenance personnel that meet the conditions, select the one with the lightest current workload for assignment. 4. Update the workload of the personnel and continue to process the next task.

[0051] The cluster constraint Latin hypercube sampling occupies 80% of the population, and the length of each operation and maintenance personnel's work order is taken as a sampling dimension, the total dimension is the number of operation and maintenance personnel m, Latin hypercube sampling is performed within the range of the number of task assignments, and uniformly distributed work order length combinations are generated to ensure orthogonal and uniform coverage of the population in the multi-dimensional search space; the sampling range of each dimension is set to , where n is the total number of tasks, and m is the total number of operation and maintenance personnel. The upper limit represents the upper limit of the number of tasks per person under average allocation.

[0052] For the s-th sample individual, the length of the j-th operation and maintenance personnel's work order is calculated as follows: , where is the permutation random number of the j-th dimension, satisfies , and for a fixed , all constitute a permutation of . is the number of individuals generated by Latin hypercube sampling, that is . .

[0053] The greedy task filling mechanism fills each operation and maintenance personnel's work order with a corresponding number of specific tasks according to the work order length determined by the Latin hypercube sampling , and the filling process selects tasks by using a comprehensive greedy score, which considers skill matching degree, clustering preference weight and geographical distance factors, and preferentially selects tasks with the highest comprehensive score and that meet the clustering constraint conditions for allocation.

[0054] Comprehensive greedy score calculation: for operation and maintenance personnel and candidate task , the comprehensive greedy score is defined as: , where is the skill matching degree function, which returns 1 when , otherwise returns 0; is the preference weight of operation and maintenance personnel for the geographical cluster to which task belongs; is the Euclidean distance between the current location of operation and maintenance personnel and the location of the task; is the weight coefficient; is a small constant to prevent division by zero.

[0055] The task filling process must satisfy the following constraints: only tasks that meet can be assigned to operation and maintenance personnel Skill constraint: the resource requirement of a task must match the skill set of an operator. Uniqueness constraint: each task can only be assigned to one operator.

[0056] Fill algorithm procedure: 1. Initialize the set of unassigned tasks 2. For each operator , repeat the following steps times: calculate the overall greedy score of all unassigned tasks for this operator; select the task with the highest score that satisfies the constraints ; assign to , remove it from ; update the current location of the operator to 3. If there are unassigned tasks, randomly select an idle operator to assign.

[0057] The orthogonal property of Latin hypercube sampling ensures that the generated individuals have good distribution characteristics in the multi-dimensional search space. Specifically, the covariance matrix of the initial population is more stable, and the propagation index of the initial Pareto front is improved.

[0058] S25, by calculating the modified local fitness of each order in each chromosome , according to the modified local fitness and the dynamic temperature parameter , a probability weighted selection mechanism is used to select parent individuals for breeding.

[0059] Specifically, in this embodiment, in order to realize the probability weighted selection, it is necessary to first decompose the global objective function into the local contribution of each operator order. For operator , its local fitness contains three components: sub-cost : the contribution of operator to the total operation cost, including labor cost and transportation cost: where is the set of tasks assigned to operator , and is the total path distance of the operator performing all assigned tasks. The path distance calculation needs to consider the distance from the starting location to the first task location, as well as the moving distance between tasks.

[0060] sub-task completion time : the total time required for operator to complete all assigned tasks: wherein, for the operation and maintenance personnel to complete the task , the time including the traffic time from the starting position or the previous task position to the task and the operation time for performing the task .

[0061] sub-response time : the weighted response time of the operation and maintenance personnel assigning the task: wherein, is the emergency degree weight of the task .

[0062] In order to prevent the over-concentration of task assignment to a few operation and maintenance personnel, a Shannon entropy regularization term based on geographical clustering is introduced. The regularization term encourages the diversified assignment of tasks among different geographical clusters.

[0063] For the operation and maintenance personnel , the task proportion of each geographical cluster in the work order of the personnel is calculated: wherein, denotes the task proportion of the geographical cluster in the work order of the operation and maintenance personnel , is the cluster label of the task , and the total number of tasks assigned to the personnel.

[0064] The Shannon entropy is calculated based on the task proportion of each cluster: wherein, is the total number of geographical clusters, is a small constant to prevent the logarithmic operation from being wrong. When the task assignment is more diversified (i.e., distributed in more different geographical clusters), the value is larger, indicating that the diversity is better.

[0065] Combining the three local objective components and the Shannon entropy regularization term, the modified local fitness is obtained: wherein, is the entropy weight coefficient, used to balance the optimization performance and the assignment diversity.

[0066] The probability of each operation and maintenance personnel being selected is calculated using the modified local fitness and the dynamic temperature parameter . wherein, is the performance of the personnel is the probability of being selected, is the dynamic temperature parameter of the current iteration, is the index of the personnel. The negative sign indicates that the smaller the fitness (i.e. the better the performance), the larger the probability of being selected.

[0067] temperature parameter controls the degree of randomness of the selection process: when is large, the selection probability tends to be uniformly distributed, increasing the randomness of the search; when is small, the selection is more biased towards individuals with good fitness, increasing the directedness of the search.

[0068] Parent selection procedure: 1. Calculate the modified local fitness of all individuals in the current population ; 2. Calculate the selection probability according to the current temperature parameter ; 3. Use the roulette wheel selection method to select parent individuals according to the probability distribution; 4. Repeat the selection process until the required number of parent individuals is obtained.

[0069] This probability-weighted selection mechanism can better utilize the fitness information of individuals compared to traditional random selection or tournament selection, while maintaining appropriate selection pressure through dynamic temperature adjustment, effectively improving the search efficiency of the algorithm and the quality of the solution.

[0070] S26, performing genetic operations to generate a child population , wherein the genetic operations include: performing a task block exchange dedicated crossover operator with a crossover probability ; and performing a 2-opt local search mutation with a mutation probability and applying a rollback mechanism.

[0071] Specifically, in this embodiment, the task block exchange dedicated crossover operator is as follows: randomly selecting a work order of an operation and maintenance personnel from parent A, and then randomly selecting a continuous task segment from the work order. Then, randomly selecting another operation and maintenance personnel from parent B that satisfies the clustering constraint (i.e. for each task in the segment, ), inserting the task segment into a random position of the work order of , while removing the tasks originally existing in the work order of to ensure that tasks are not assigned repeatedly.

[0072] The 2-opt local search mutation operator is as follows: Randomly select a work order from a specific chromosome for an operations and maintenance personnel, and apply the 2-opt algorithm to this task list for local path optimization. During the iteration of 2-opt, if the mutated local fitness... Then, based on the rollback probability Roll back to the old state to avoid degradation. For the local fitness after mutation, The local fitness before mutation, This represents the current iteration number. This represents the maximum number of iterations.

[0073] S27, By merging parent populations and offspring population Forming a joint population Crowding distance is calculated for the joint population, and non-dominated ranking is performed to select superior individuals to form the next generation of the population. .

[0074] Specifically, in this embodiment, the population merging process maintains complete information about all individuals, including attributes such as chromosome encoding, objective function value, and local fitness.

[0075] For any two individuals in the joint population and ,individual Dominant Individual (recorded as) If and only if: in, These are the objective functions for total maintenance cost, total task completion time, and response time, respectively.

[0076] The joint population is stratified using a fast non-dominated sorting algorithm: First floor ( ): Identify all non-dominated individuals, i.e., individuals not dominated by any other individuals; Subsequent layer construction: Continue searching for non-dominated individuals from the remaining individuals until all individuals are assigned to the appropriate layer; Layer labeling: Assign a non-dominated level to each individual. , Individuals in , Individuals in And so on.

[0077] Crowded distance Used to measure individuals The density in the target space is calculated by the formula: wherein, and are the maximum and minimum values of the target in the current layer, respectively. The adjacent individuals on the th target, and are the maximum and minimum values of the target in the current layer, respectively.

[0078] The calculation steps are described as follows: for each non-dominated layer , the crowding distance of each individual in the layer is calculated; for each objective function , the individuals in the layer are arranged in ascending order of value; the crowding distances of the first and last individuals in the sorted order are set to infinity; for the intermediate individuals, the local crowding distances of the individuals on each target are accumulated.

[0079] From the joint population , select individuals to form the next generation population according to the following priority order: prefer individuals with smaller values; among individuals with the same value, prefer individuals with larger values.

[0080] S28, determine whether the iteration stop condition is reached, if yes, output the current Pareto optimal solution set, if not, iteration number t = t + 1, update the dynamic temperature parameter , and return to step S25 to continue iteration.

[0081] Specifically, in this embodiment, the iteration stop condition is to reach the maximum iteration number or to reach the convergence condition, wherein the convergence condition is that when the change rate of the non-dominated solution set within the continuous observation window is less than the convergence threshold, it is determined to be converged.

[0082] The temperature parameter adopts a dynamic adjustment strategy based on Logistic chaotic mapping: wherein, is a chaotic parameter, so that the system exhibits chaotic characteristics; is the temperature parameter of the th generation, and the initial value is ; is a disturbance amplitude, and a periodic disturbance is introduced; is the current iteration number, is the maximum iteration number.

[0083] To prevent the temperature parameter from deviating too much from the reasonable range, a boundary constraint is set: It should be noted that the values of the upper and lower bounds in the boundary constraint are empirical values. The lower bound of 0.1 ensures that the selection mechanism maintains a minimum degree of randomness, and the upper bound of 50.0 ensures that the algorithm can still use fitness information for biased selection in the initial and perturbation stages. In addition, when the actual value is taken, the upper bound value should match the numerical range of the local fitness, i.e. The calculation is reasonable.

[0084] When the algorithm meets the termination condition, the first non-dominated layer is extracted from the current population as the Pareto optimal solution set: If the termination condition is not met, the following update operations are performed: the iteration counter is incremented: ; the temperature parameter is updated: the temperature update formula is applied. Return to step S25 to start a new round of probability weighted selection process, forming a complete evolution cycle.

[0085] Specifically, in an embodiment of the present application, step S3 selects the final execution scheme from the Pareto optimal solution set and converts it into an executable format, outputting the complete operation and maintenance scheduling scheme.

[0086] The final scheme selection includes automatic selection mode and manual selection mode: Automatic selection mode: automatically select the optimal scheme from the solution set according to the preset preference. Calculate the comprehensive score of each solution by a multi-objective decision function, and select the solution with the highest comprehensive score as the final execution scheme. The decision function considers the weighted combination of three objectives: total operation and maintenance cost F1, total task completion time F2, and response time F3, and the weights can be configured according to actual operation and maintenance strategy requirements.

[0087] Manual selection mode: present the Pareto front to the operation supervisor in a visual manner for manual selection. The visual interface displays a three-dimensional scatter plot, with each point representing a Pareto optimal solution and the three coordinate axes corresponding to the three optimization objectives. The operation supervisor can view the detailed information of each solution and make a decision based on the current operation strategy and resource status.

[0088] After determining the final scheme, convert the chromosome encoding format to a standardized execution scheme: Personnel scheduling list: generate an independent work task list for each operation and maintenance personnel , including personnel basic information, assigned task list , task detailed information (location coordinates , emergency level weight , estimated working hours , resource requirements , and cluster assignment information.

[0089] Path planning instructions: Calculate the optimal navigation path based on the task sequence of each maintenance personnel. From the starting position of the maintenance personnel , generate standardized navigation instructions containing GPS coordinates, estimated driving distance and time according to the location coordinates in the task sequence.

[0090] Material and tool list: According to the resource requirements of all assigned tasks , generate a material list. The system aggregates the resource requirements of all tasks, removes duplicates and generates a personal material list for each maintenance personnel.

[0091] Overall time schedule: Create a Gantt chart format schedule containing all task time schedules. Calculate the start time, duration , and completion time for each task , generate an independent timeline for each maintenance personnel, showing the task execution order and time schedule.

[0092] In addition, as Figure 3 shown, the present application also provides a photovoltaic power generation system operation and maintenance management platform, which is used to realize the method of any one of the above, and the platform comprises: A data acquisition module is used to acquire photovoltaic equipment running state data, fault alarm information, maintenance personnel position information and skill resource data in real time; the module is connected with the photovoltaic monitoring system, GPS positioning system and human resource management system through the interface, and acquires photovoltaic equipment running state data (such as power generation power, equipment temperature, current and voltage), fault alarm information (including fault type, position coordinates and emergency degree), maintenance personnel position information (starting position coordinates ) and skill resource data (skill set hourly labor cost ) in real time. The module adopts a standardized data interface protocol to ensure the accuracy and real-time performance of the data.

[0093] A task management module is used to receive maintenance task requirements, perform task classification, priority evaluation and resource requirement analysis; the module receives fault alarms and regular maintenance requirements from the data acquisition module, and automatically generates a maintenance task set . Each task contains geographic location coordinates , emergency degree weight , estimated working hours and resource requirements The module classifies tasks and evaluates priorities according to fault types, influence ranges and historical experience, and provides standardized task data for subsequent intelligent scheduling.

[0094] The intelligent scheduling module is used for multi-objective optimization calculation based on an improved NSGA-II algorithm to generate a Pareto optimal solution set; the module executes the complete algorithm process of step S2, including parameter initialization, k-means geographical clustering, cluster-enhanced coding, hybrid population initialization, probability weighted selection, genetic operation and environment selection, etc. The module uses a parallel computing architecture to improve the algorithm execution efficiency, and processes large-scale task scheduling problems through multi-threading. Finally, a Pareto optimal solution set containing multiple non-dominated solutions is output, providing decision makers with diversified scheduling scheme selection.

[0095] The decision support module is used for providing two modes of automatic selection and manual selection to determine the final execution scheme from the Pareto optimal solution set; the automatic selection mode automatically calculates a comprehensive score according to a preset preference weight to select an optimal scheme; the manual selection mode displays a Pareto front through a three-dimensional visualization interface, each solution is displayed in the form of a scatter point with three target values of total operation and maintenance cost, total task completion time and response time, and operation and maintenance supervisors can interactively view detailed information of the solution and manually select.

[0096] The execution monitoring module is used for outputting a complete operation and maintenance execution scheme, including a personnel scheduling list, path planning instructions, a material tool list and a time schedule, and monitoring the execution state in real time.

[0097] The data storage module is used for storing historical operation and maintenance data, optimization parameter configuration and scheduling scheme records. The module stores historical operation and maintenance data for experience accumulation and data mining, optimization parameter configuration information for algorithm tuning, and scheduling scheme records for effect evaluation and continuous improvement.

[0098] The above only describes the preferred embodiments of the present application and does not limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for operation and maintenance management of a photovoltaic power generation system, characterized by, Comprise: S1, obtaining a set of operation and maintenance tasks to be processed and a set of idle operation and maintenance personnel wherein each operation and maintenance task contains geographical position coordinates, emergency degree weight, estimated work hours and resource requirements, and each operation and maintenance personnel contains a set of skills, starting position coordinates and hourly labor cost; is the total number of operation and maintenance tasks, is the total number of operation and maintenance personnel; S2, the operation and maintenance task is scheduled optimization through improved NSGA-II algorithm, the multi-objective optimization problem of total operation and maintenance cost minimization, total task completion time minimization and response time minimization is established, clustering enhanced coding structure, hybrid population initialization strategy and probability weighted genetic operator are used to solve multi-objective optimization, and the Pareto optimal solution set is output; S3, by selecting a final scheme from the Pareto optimal solution set and converting into executable format, output the complete operation and maintenance scheduling scheme including personnel scheduling list, path planning instruction, material and tool list and overall time schedule. 2.The operation and maintenance management method of a photovoltaic power generation system according to claim 1, characterized in that, Step S2 comprises: S21, set population size , maximum number of iterations , crossover probability , mutation probability , initial temperature parameter , number of clusters K and cluster preference threshold ; S22. Use the k-means algorithm to cluster the geographic coordinates of all operation and maintenance tasks to obtain K geographic clusters, and then cluster each operation and maintenance task. Assign cluster labels The clustering preference weight vector is calculated based on the Euclidean distance between the starting position of the maintenance personnel and each cluster center. A clustering-enhanced coding scheme is used to define chromosome X as... ,in To be allocated to maintenance personnel Task sequence, For this person The clustering preference weight vector is used, and clustering constraints are set to limit task allocation; S23, by establishing the multi-objective optimization function of total operation and maintenance cost minimization, total task completion time minimization and response time minimization, the optimization target is determined; S24, generating an initial population by a mixed population initialization strategy Wherein 20% of the individuals are generated by cluster-constrained greedy seed generation, 80% of the individuals are generated by cluster-constrained Latin hypercube sampling, and the iteration number t is set to 0. S25. By calculating the corrected local fitness of each work order in each chromosome. Based on the modified local fitness and dynamic temperature parameters A probability-weighted selection mechanism is used to select parent individuals for reproduction; S26, performing a genetic operation to generate an offspring population wherein the genetic operation comprises: with a cross probability a dedicated cross operator performing task block exchange; with a mutation probability Perform 2-opt local search mutation and apply rollback mechanism; S27, merging the parent population with the offspring population to form a combined population S28, performing crowding distance calculation and non-dominated sorting on the combined population to select elite individuals to form the next generation population S29, repeating steps S21 to S28 until a termination condition is met ; S28, judging whether the iteration stopping condition is reached, if yes, outputting the current Pareto optimal solution set, if not, iteration number t = t + 1, updating the dynamic temperature parameter and returning to step S25 to continue iteration. 3.The operation and maintenance management method of a photovoltaic power generation system according to claim 2, characterized in that, Step S22 comprises: Clustering task locations using a k-means algorithm to obtain K geographic clusters, each task being assigned a cluster label ; wherein the cluster centers are computed: ​ In the formula, Let k be the cluster center of the k-th geographical cluster. Let k be the number of tasks in the k-th geographic cluster. For the task Geographical coordinates; For each operations personnel Computing a preference weight vector for each geographical cluster : wherein is the start position coordinate of the maintenance personnel is the preference weight vector of the geographic cluster k, is the Euclidean distance function, is the start position coordinate of the maintenance personnel is the start position coordinate of the maintenance personnel is a small constant to prevent division by zero; The normalized processing obtains the maximum weight vector: In the formula, K is the total number of clusters, is the cluster index; Setting cluster constraints: task Only if the cluster preference condition is met, the task can be assigned to a person where is the cluster preference threshold, is the cluster label of the task, and denotes the th component of the weight vector , i.e. . 4.The operation and maintenance management method of a photovoltaic power generation system according to claim 2, characterized in that, Multi-objective optimization function comprises: Total operation and maintenance cost minimization: In the formula, represents the chromosome corresponding to the total operation and maintenance cost, is the total number of operation and maintenance personnel, is the operation and maintenance personnel hourly labor cost, is the task set assigned to the operation and maintenance personnel , is the task estimated work hours, is the unit distance transportation cost, is the total path distance of the operation and maintenance personnel ; Total task completion time minimization: In the formula, represents a chromosome corresponding total task completion time, is the total number of operation and maintenance tasks, is the completion time of the task . Response time minimization: In the formula, Chromosomes The corresponding total response time, For the task The urgency level weight. 5.The operation and maintenance management method of a photovoltaic power generation system according to claim 2, characterized in that, Hybrid population initialization strategy comprises: Cluster constraint greedy seed generation accounts for 20% of the population, including cost priority seed and urgency priority seed, cost priority seed allocates each task to the nearest operation and maintenance personnel who meets the clustering constraint condition, and urgency priority seed allocates tasks to idle operation and maintenance personnel in turn according to task urgency from high to low, which meets the clustering constraint condition; Cluster constraint Latin hypercube sampling accounts for 80% of the population, the length of each operation and maintenance personnel's work order is taken as a sampling dimension, the total dimension is the number of operation and maintenance personnel m, Latin hypercube sampling is carried out within the range of task allocation number, and uniform distribution of work order length combination is generated to ensure orthogonal uniform coverage of the population in the multi-dimensional search space; The greedy task filling mechanism fills the work order of each operation and maintenance personnel with the corresponding number of specific tasks according to the work order length determined by Latin hypercube sampling, and the filling process selects tasks according to the comprehensive greedy score, which considers skill matching degree, clustering preference weight and geographical distance factors, and preferentially selects the task with the highest comprehensive score and meeting the clustering constraint condition for allocation.

6. The operation and maintenance management method of a photovoltaic power generation system according to claim 4, characterized in that, In step S25, the probability weighted selection mechanism is specifically: Calculate the modified local fitness of each work order: In the formula, is the modified local fitness of the operation and maintenance personnel , is the sub-cost of the operation and maintenance personnel , is the sub-task completion time of the operation and maintenance personnel , is the sub-response time of the operation and maintenance personnel , is the entropy weight coefficient, is the Shannon entropy regularization term; Calculate the selection probability by using softmax function: In the formula, for the operation and maintenance personnel the probability of being selected, is a dynamic temperature parameter of the current iteration, is a personnel index.

7. The operation and maintenance management method of a photovoltaic power generation system according to claim 6, characterized in that, In step S26: The special crossover operator adopts task block exchange mode, selects the continuous task fragments of operation and maintenance personnel's work order from parent A, and inserts them into the operation and maintenance personnel's work order that meets the clustering constraint in parent B; The 2-opt local search mutation operator optimizes the path of the task sequence and employs a rollback mechanism: when the local fitness is mutated... At that time, based on the rollback probability Rollback to the old state, where For the local fitness after mutation, The local fitness before mutation, This represents the current iteration number. This represents the maximum number of iterations. 8.The operation and maintenance management method of a photovoltaic power generation system according to claim 2, characterized in that, The iteration stopping condition is reaching a maximum number of iterations or reaching a convergence condition, wherein the convergence condition is that convergence is determined when a rate of change of the non-dominated solution set within a consecutive observation window is less than a convergence threshold. 9.The operation and maintenance management method of a photovoltaic power generation system according to claim 1, characterized in that, In step S3, selecting the final execution scheme from the Pareto optimal solution set comprises two modes: Automatic selection mode automatically selects the optimal scheme from the solution set according to the preset preference; Manual selection mode presents the Pareto front to the operation and maintenance supervisor in a visual way for manual selection.

10. An operation and maintenance management platform of a photovoltaic power generation system, characterized in that, The platform is used to realize the method of any one of claims 1-9, and the platform comprises: A data acquisition module for acquiring real-time photovoltaic equipment operation state data, fault alarm information, operation and maintenance personnel location information and skill resource data; A task management module for receiving operation and maintenance task requirements, performing task classification, priority evaluation and resource demand analysis; An intelligent scheduling module is configured to perform multi-objective optimization calculation based on an improved NSGA-II algorithm to generate a Pareto optimal solution set; A decision support module is configured to provide two modes of automatic selection and manual selection to determine a final execution scheme from the Pareto optimal solution set; An execution monitoring module is configured to output a complete operation and maintenance execution scheme, including a personnel scheduling list, path planning instructions, a material tool list and a time schedule, and to monitor the execution status in real time; A data storage module is configured to store historical operation and maintenance data, optimization parameter configurations and scheduling scheme records.

Images