CN-121981370-A - Fan address selection method, device, equipment and storage medium

CN121981370ACN 121981370 ACN121981370 ACN 121981370ACN-121981370-A

Abstract

The embodiment of the disclosure provides a fan position selecting method, device, equipment and storage medium, wherein the method comprises the steps of determining boundary conditions of a position selecting area, rasterizing the position selecting area, determining a feasible area and an infeasible area in the position selecting area according to the boundary conditions, deleting grid points which are not in the feasible area or in the infeasible area to obtain remaining grid points and drawing a grid map, acquiring local wind resource data, formulating an objective function according to a preset target, calculating an objective function value of each grid point in the grid map based on the local wind resource data and the objective function, randomly generating an adoption point on the grid map by utilizing the boundary conditions, generating an initial population according to the adoption point, taking the initial population as a parent population in first iteration, circulating iteration, outputting a final child population, and further determining the fan position. The problem of frequent feasibility verification caused by taking the number of fans as the constraint of an equation is avoided, the searching efficiency of a genetic algorithm is remarkably improved, and the computing resource is saved.

Inventors

LIU KUN
HAN SHUWEI
LI XINYU
LIU PENGPENG
SONG YIPENG
LIU GEFEI
ZHOU ZHONGCHENG

Assignees

中国三峡新能源(集团)股份有限公司
三峡新能源发电(安丘市)有限公司
中国长江三峡集团有限公司

Dates

Publication Date: 20260505
Application Date: 20251222

Claims (10)

1. A method for locating a blower, comprising: Determining boundary conditions of the addressing area, rasterizing the addressing area, determining a feasible area and an infeasible area in the addressing area according to the boundary conditions, deleting grid points which are not in the feasible area or in the infeasible area to obtain residual grid points, and drawing a grid map; Acquiring local wind resource data, formulating an objective function according to a preset target, and calculating an objective function value of each grid point in a grid map based on the local wind resource data and the objective function, wherein the preset target is to enable the sum of wind speed and module length of each machine position to be maximum; Randomly generating an adopted point on a grid map by utilizing the boundary condition, and generating an initial population according to the adopted point; When iteration is performed for the first time, an initial population is taken as a parent population, a selection operator is applied to the parent population, a quasi-intersecting individual is determined, an improved intersecting operator is applied to the quasi-intersecting individual, a quasi-offspring individual is determined, the selection process of the selection operator and the intersecting process of the improved intersecting operator are repeated until the obtained quantity of the quasi-offspring individuals is the same as that of the initial individuals, a quasi-offspring population is determined, an improved mutation operator is applied to the quasi-offspring population, a mutation quasi-offspring population is determined, the parent population and the mutation quasi-offspring population are combined, the combination result is screened according to an objective function value corresponding to the combination result, a offspring population is obtained, the offspring population generated each time is taken as the parent population of the next iteration, iteration is circulated until a preset termination criterion is met, a final offspring population is output, and the final offspring population is utilized to determine the fan machine position.
2. The method of claim 1, wherein determining boundary conditions of the addressing area while rasterizing the addressing area, determining viable and non-viable areas within the addressing area based on the boundary conditions, deleting grid points that are not in the viable or non-viable areas to obtain remaining grid points, and drawing a grid map, comprises: Rasterizing the address selecting area according to preset grid precision; Determining a boundary condition according to a preset requirement, and determining a first constraint condition, a second constraint condition and a third constraint condition according to the boundary condition; Determining a feasible region and an infeasible region in the site selection region according to the first constraint condition and the second constraint condition; Verifying each grid point, deleting the grid points which are not in a feasible area or in an infeasible area, and obtaining the rest grid points; and renumbering the rest grid points, and drawing a grid map of the addressing area according to the renumbered grid points.
3. The method of claim 1, wherein obtaining local wind resource data, formulating an objective function based on a preset objective, and calculating an objective function value for each grid point in a grid map based on the local wind resource data and the objective function, comprises: acquiring local wind resource data, and generating a wind resource distribution map aligned with a grid map by utilizing the local wind resource data; and formulating an objective function according to a preset objective, and calculating an objective function value of each grid point in the grid map based on the wind resource distribution diagram and the objective function.
4. The method of claim 2, wherein randomly generating the adoption points on the grid map using the boundary condition and generating the initial population from the adoption points comprises: the grid map comprises row vectors with 0 elements which are the same as the number of the rest grid points; Determining a first preset range according to the number of the remaining grid points, generating random integers in the first preset range in a uniform distribution mode, and verifying whether the distances between the grid points corresponding to the random integers and all the grid points adopted in the row vector meet a third constraint condition or not; If yes, changing an element corresponding to the random integer in the row vector to be 1; if the number of the elements in the row vector is not equal to the number of the preset machine bits, regenerating a random integer, and continuing to verify whether the distances between the grid point corresponding to the random integer and all the grid points adopted in the row vector meet a third constraint condition or not until the number of the elements in the row vector which are 1 is equal to the number of the preset machine bits, and outputting the row vector as an initial individual; repeating the steps for generating a plurality of initial individuals for a plurality of times, and obtaining an initial population according to all the initial individuals.
5. The method of claim 4, wherein upon a first iteration, using the initial population as a parent population, applying a selection operator to the parent population, determining a proposed cross-over individual, applying an improved cross-over operator to the proposed cross-over individual, determining a proposed sub-population, repeating the selection process of the selection operator and the cross-over process of the improved cross-over operator until the number of proposed sub-population is the same as the number of initial individuals, and determining a proposed sub-population, comprising: In the first iteration, taking the initial population as a parent population; Selecting a first parent and a second parent from the parent population through a selection operator, applying an improved crossover operator to the first parent and the second parent, determining the difference between the first parent and the second parent, classifying the difference into two types, and obtaining a first type and a second type, wherein the number of the difference between the first type and the second type is equal; Obtaining a second preset range according to the number of the differences, generating difference random integers in the second preset range in a uniform distribution mode, randomly selecting columns of the number of the difference random integers in the first class and the second class according to the uniform distribution mode, and obtaining a selection result; Changing the corresponding element of the first parent first class from 1 to 0, changing the corresponding element of the first parent second class from 0 to 1, changing the corresponding element of the second parent first class from 0 to 1, changing the corresponding element of the second parent second class from 1 to 0, generating a first pseudo-offspring according to the first parent in the change result, and generating a second pseudo-offspring by the second parent; Judging whether the first simulated offspring and the second simulated offspring meet a third constraint condition, and if the first simulated offspring and the second simulated offspring meet the third constraint condition, selecting a new first father and a new second father; If only one of the first and second quasi-offspring accords, saving the matched quasi-offspring, and simultaneously, reapplying an improved crossover operator to the first and second father to obtain a new first quasi-offspring and a new second quasi-offspring, judging whether the new first and second quasi-offspring accords with a third constraint condition, and if the new first and second quasi-offspring accords with the third constraint condition, randomly selecting a quasi-offspring individual to save; And repeating the selection process of the selection operator and the crossing process of the improved crossing operator until the number of the obtained pseudo-offspring individuals is the same as the number of the initial individuals, and determining the pseudo-offspring population.
6. The method of claim 5, wherein applying a selection operator to the parent population to determine individuals to cross comprises: Determining the objective function values corresponding to all the parent individuals in the parent population according to the objective function values of each grid point in the grid map, and calculating the corresponding selected probability and the accumulated selected probability according to the objective function values corresponding to each parent individual in the parent population; Generating a plurality of first probability random numbers in a third preset range in a uniform distribution manner, forming an inequality set according to the accumulated selected probabilities corresponding to two adjacent individuals, mapping each probability random number into a corresponding inequality to be met, and forming a inequality set to be met, wherein the inequality to be met is that the probability random number meets the left side of the inequality and the right side of the inequality at the same time; and taking the parent individuals corresponding to the accumulated selected probabilities on the right side of each inequality in the inequality set as the cross-fitting individuals.
7. The method of claim 1, wherein applying the modified mutation operator to the simulated offspring population, determining the mutated simulated offspring population, combining the parent population with the mutated simulated offspring population, screening the combined result according to the objective function value corresponding to the combined result to obtain the offspring population, taking each generated offspring population as the parent population of the next iteration, iterating circularly until a preset termination criterion is met, outputting the final offspring population, and determining the fan position by using the final offspring population, comprising: Applying the improved mutation operator to the pseudo-offspring population, uniformly distributing the second probability random number in a third preset range, and if the second probability random number is not larger than the preset mutation probability, generating a first digit random integer in a fourth preset range, and uniformly distributing the second digit random number in a fifth preset range; The fourth preset range is obtained according to the number of total machine digits, and the fifth preset range is obtained according to the difference value between the number of remaining grid points and the number of total machine digits; Randomly mutating the pseudo-offspring individuals in the pseudo-offspring population, changing the first digit random integer number of elements 1 of the pseudo-offspring individuals into 0, and changing the second digit random integer number of elements 0 into 1 to obtain mutated pseudo-offspring; If the variant offspring accords with the third constraint condition, storing; if the variant offspring does not meet the third constraint condition, the corresponding offspring individual is subjected to re-variation until the variant offspring individual meets the third constraint condition and is stored, and a variant offspring population is comprehensively obtained; Merging the parent population and the variant simulated offspring population based on elite strategy, calculating an objective function value of each element in the merging result, sorting each element in the merging result from big to small according to the objective function value, and reserving the upper half part of the sorting result to obtain the offspring population; And taking the child population generated each time as the parent population of the next iteration, performing loop iteration until a preset termination criterion is met, outputting a final child population, and determining the fan position by using the final child population.
8. A blower location apparatus, the apparatus comprising: The first determining module is used for determining boundary conditions of the addressing area, rasterizing the addressing area at the same time, determining a feasible area and an infeasible area in the addressing area according to the boundary conditions, deleting grid points which are not in the feasible area or in the infeasible area to obtain residual grid points, and drawing a grid map; The formulating module is used for obtaining local wind resource data, formulating an objective function according to a preset target, and calculating an objective function value of each grid point in the grid map based on the local wind resource data and the objective function, wherein the preset target is to enable the sum of wind speed module lengths of all machine positions to be maximum; the generation module is used for randomly generating an adopted point on the grid map by utilizing the boundary condition and generating an initial population according to the adopted point; The second determining module is used for applying a selection operator to a parent population to determine a quasi-intersecting individual when the first iteration is performed, applying an improved intersecting operator to the quasi-intersecting individual, determining a quasi-offspring individual, repeating the selection process of the selection operator and the intersecting process of the improved intersecting operator until the number of obtained quasi-offspring individuals is the same as the number of the initial individuals, determining a quasi-offspring population, applying an improved mutation operator to the quasi-offspring population, determining a mutation quasi-offspring population, combining the parent population and the mutation quasi-offspring population, screening the combination result according to the objective function value corresponding to the combination result to obtain a offspring population, taking the offspring population generated each time as the parent population of the next iteration, performing loop iteration until a preset termination criterion is met, outputting a final offspring population, and determining the fan position by using the final offspring population.
9. An electronic device, comprising: A memory; processor, and A computer program; Wherein the computer program is stored in the memory and configured to be executed by the processor to implement the blower location selection method of any one of claims 1-7.
10. A computer readable storage medium, characterized in that the storage medium stores a computer program which, when executed by a processor, causes the processor to implement the method of providing blower addressing of any one of claims 1-7.

Description

Fan address selection method, device, equipment and storage medium Technical Field The disclosure relates to the technical field of wind power layout, in particular to a fan position selecting method, a device, equipment and a storage medium. Background In the context of large-scale development of land-based wind power, the work of microscopic machine site selection is facing new practical constraints. On one hand, land resources suitable for building wind power plants are increasingly tense, the number of areas which are flat and open, have good wind resources and have access conditions is continuously reduced, on the other hand, the requirements on ecological protection red lines, environmental constraint, resident safety distance, landscape protection and the like are increasingly strict, the selectable machine position space is cut in a large quantity, and the freedom degree of fan arrangement is obviously reduced. By superposition of all factors, the micro site selection difficulty of the wind power project is remarkably increased, and the method becomes one of key bottlenecks for limiting continuous and high-quality development of onshore wind power. In order to improve the microscopic site selection efficiency of the fan, the prior art generally carries out systematic optimization layout on the fan position by establishing a mathematical optimization model, and adopts an intelligent optimization algorithm to solve. Wherein, intelligent algorithms based on random search such as genetic algorithm, particle swarm algorithm, etc. are widely applied to the fan position optimization problem. In order to ensure that the number of fans in the optimization result is consistent with the project stand or the agreed installed capacity in the planning file, the existing model often writes the 'fixed number of fans' as an equality constraint into the optimization model. However, the fan addressing problem is essentially a highly nonlinear, strong constraint and discrete combined optimization problem, which is to satisfy complex space constraints (such as feasible region/infeasible region boundaries, minimum safety spacing between machine positions and the like) and also consider factors such as wind resource distribution, fan wake effect, electrical access and the like. In this context, when the number of fans is explicitly added to the model by means of equality constraint, for the genetic algorithm class method based on random search, a strict feasibility check must be performed on new solutions (individuals) generated in each generation and each iteration to determine whether the number of fans meets both the number of fans and the space constraint. If not, the solution needs to be discarded or repaired, which can cause the following problems. First, the feasibility check is frequent. Under the conditions of large population scale and more iteration times, a large number of crossed and mutated individuals can be judged to be infeasible solutions because the number constraint of fans is not met, and the algorithm needs to frequently execute feasibility checking and repairing operations, so that the overall calculation burden is increased. Second, computing resources are wasted. The generation and inspection of a large number of infeasible solutions not only occupy computation time, but also occupy a large amount of storage and operation resources, and the computation time is particularly remarkable for the case of large-scale wind farms or high-precision grid division. In summary, in the prior art, when a random search type algorithm such as a genetic algorithm is used to solve the fan position optimization problem, there is a technical problem that the feasibility verification is frequent and the computing resource consumption is relatively large, and a new site selection modeling and solving method is necessary to be provided, and on the premise that the number of fans and space constraints are met, the dependence on equality constraint feasibility check is reduced, so that the searching efficiency of the genetic algorithm is improved and the computing cost is reduced. Accordingly, embodiments of the present disclosure provide a method, an apparatus, a device, and a storage medium for fan address selection. Disclosure of Invention In order to solve the technical problems described above or at least partially solve the technical problems described above, the present disclosure provides a method, an apparatus, a device and a storage medium for fan location selection. The embodiment of the disclosure provides a fan position selecting method, which comprises the following steps: Determining boundary conditions of the addressing area, rasterizing the addressing area, determining a feasible area and an infeasible area in the addressing area according to the boundary conditions, deleting grid points which are not in the feasible area or in the infeasible area to obtain residual grid points, and drawing a grid map; Ac