Agricultural photovoltaic array multi-objective collaborative optimization method based on light energy coupling regulation

Abstract

This invention relates to a multi-objective collaborative optimization method for agricultural photovoltaic arrays based on light-energy coupling regulation. The method includes acquiring basic information data for agricultural photovoltaic array layout optimization; obtaining canopy light distribution data, low light ratio, and continuous shading duration; correcting zoning errors; establishing a multi-objective optimization model; establishing and validating a response surface model; obtaining a Pareto candidate layout solution set; selecting the agricultural photovoltaic array layout scheme with the best overall performance; determining whether the verification error meets a preset threshold; and selectively outputting the optimal layout parameters or supplementing sampling and reconstructing the response surface model. The multi-objective collaborative optimization method for agricultural photovoltaic arrays based on light-energy coupling regulation designed in this invention can solve the technical problem of the difficulty in coordinating the optimization of crop light intensity, power generation, and light uniformity throughout the entire growth period in agricultural photovoltaic array layout.

Description

Technical Field

[0001] This invention relates to the field of agricultural photovoltaic system layout optimization technology, specifically to a multi-objective collaborative optimization method for agricultural photovoltaic arrays based on light energy coupling regulation. Background Technology

[0002] Agricultural photovoltaic (PV) systems utilize photovoltaic (PV) modules deployed within agricultural planting areas, enabling the same land space to serve both crop production and PV power generation functions. The array arrangement of such systems not only affects the PV modules' ability to receive solar radiation but also alters the spatial distribution of sunlight on the farmland surface and under the crop canopy. Different crop types and growth stages exhibit variations in canopy height, light requirements, shading sensitivity, and space requirements for agricultural machinery operations; therefore, the layout of agricultural PV arrays needs to be coordinated between agricultural production conditions and PV engineering requirements.

[0003] Current agricultural photovoltaic array designs typically rely primarily on engineering experience, fixed spacing, tilt angle of individual modules, or single power generation targets, rarely considering the coupling relationship between canopy height variations throughout the crop's growth cycle, low light intensity, duration of continuous shading, light uniformity, agricultural machinery access space, and power generation per unit area. This approach can easily result in candidate layouts meeting requirements in a particular season or growth period, but exhibiting problems such as insufficient canopy light reception, prolonged localized shading, or decreased land power generation efficiency during other critical growth periods.

[0004] Therefore, existing technologies have shortcomings and need to be improved and developed. Summary of the Invention

[0005] The present invention provides a multi-objective collaborative optimization method for agricultural photovoltaic arrays based on light energy coupling regulation, which is used to solve the technical problem that it is difficult to coordinate the optimization of crop illumination, power generation and illumination uniformity throughout the entire growth period in the layout of agricultural photovoltaic arrays in the prior art.

[0006] This invention provides a multi-objective collaborative optimization method for agricultural photovoltaic arrays based on light energy coupling regulation, comprising:

[0007] Acquire basic information data for optimizing the layout of agricultural photovoltaic arrays. The basic information data includes geographical and meteorological data of the target construction area, solar radiation data, photovoltaic module structural parameters, farmland spatial parameters, crop type, growth period data, canopy height parameters, light requirement threshold, and shading sensitivity coefficient.

[0008] Based on the aforementioned basic information data, an agricultural photovoltaic system light environment simulation model was established, and a canopy calculation plane was established according to the canopy height at each growth stage. The unshaded control area at the same time was used as the normalization benchmark to obtain canopy light distribution data, low light ratio, and continuous shading duration under each combination of layout parameters.

[0009] Based on the measured errors in the area below the photovoltaic modules, the area between the photovoltaic modules, and the unshaded area, the canopy illumination distribution data is corrected for partitioning errors.

[0010] Using the tilt angle of photovoltaic modules, the installation height of modules, the array span, and the lateral gap of modules as optimization variables, a multi-objective optimization model is established, which includes crop canopy irradiance, photovoltaic power generation per unit area, and irradiance uniformity.

[0011] A response surface model was established and validated based on the samples after partitioning error correction.

[0012] The non-dominated sorting genetic algorithm II is used to perform a global search in the feasible layout parameter space to obtain an initial Pareto solution set. The initial Pareto solution set is then used as the initial fish swarm of the artificial fish swarm algorithm for local updates to obtain a Pareto candidate layout solution set.

[0013] The Pareto candidate layout solution set that meets the light requirement threshold, low light ratio, continuous shading duration and agricultural machinery passage constraints of the entire crop growth period is comprehensively evaluated by the entropy weight method and the approximate ideal solution ranking method, and the agricultural photovoltaic array layout scheme with the best comprehensive performance is selected.

[0014] Based on the light environment simulation verification data and field test verification data of the optimal agricultural photovoltaic array layout scheme, determine whether the verification error meets the preset threshold. If it does, output the optimal layout parameters; otherwise, supplement the sampling and reconstruct the response surface model.

[0015] Furthermore, the step of establishing a simulation model of the light environment of the agricultural photovoltaic system based on the basic information data, and establishing a canopy calculation plane according to the canopy height at each growth stage, includes:

[0016] Meteorological radiation data is cleaned and time-scaled, photovoltaic module structural parameters are converted, farmland space is modeled in three dimensions or meshed, and crop canopy calculation plane is discretized.

[0017] Establish a simulation model of the light environment of an agricultural photovoltaic system, including a three-dimensional structural model of photovoltaic modules, a farmland surface model, a crop canopy calculation plane, and solar radiation boundary conditions;

[0018] The entire growth period of the target crop is divided into several growth periods, and the first... Canopy height during each reproductive period or canopy height function and according to the canopy height or canopy height function Establish a crop canopy computational plane ,in This refers to the reproductive period number. For evaluation time steps;

[0019] Calculation planes of each crop canopy The grid is discretized according to the planting row spacing, plant spacing, component projection boundary and agricultural machinery passage belt, and used to extract the illuminance of each position inside the agricultural photovoltaic array relative to the unshaded control area at the same time.

[0020] No. The first crop canopy calculation plane at each growth stage Each sampling point at time... Illuminance is determined according to the following formula: In the formula, For the first The first crop canopy calculation plane at each growth stage Each sampling point at time... Illuminance; The sampling point number; For the first agricultural photovoltaic array The reproductive period, the first Each sampling point, time The intensity of solar radiation or light intensity; The solar radiation intensity or light intensity of the unobstructed control area at the same time.

[0021] Furthermore, the step of correcting the canopy illumination distribution data for zoning errors based on the measured errors in the area below the photovoltaic modules, the area between the photovoltaic modules, and the unshaded area includes:

[0022] Acquire measured solar radiation data from solar radiation sensors deployed in the area below the photovoltaic modules, the area between the photovoltaic modules, and the unshaded control area;

[0023] The measured solar radiation data are compared with the output results of the agricultural photovoltaic system light environment simulation model, and the relative error is calculated according to the area below the photovoltaic module, the area between the photovoltaic modules, and the unshaded area.

[0024] Based on the relative error of each partition, a partition error correction coefficient is established, and the canopy illumination distribution data output by the simulation is corrected using the partition error correction coefficient.

[0025] When the relative error of any partition is greater than 8%, the supplementary sample points of the corresponding partition are added to the sample set and the response surface training points are regenerated until the corrected error meets the requirements of subsequent optimization calculations.

[0026] Furthermore, the multi-objective optimization model, which uses the tilt angle of photovoltaic modules, module installation height, array span, and lateral gap of modules as optimization variables, and includes crop canopy irradiance, photovoltaic power generation per unit area, and irradiance uniformity, includes:

[0027] Set the tilt angle of the photovoltaic module to be The component installation height is The array span is The lateral gap between components is This forms the optimization variable vector: In the formula, Optimize the variable vector for agricultural photovoltaic array layout;

[0028] The optimized variable vector The following constraints must be satisfied: In the formula, and These are the lower limit and upper limit of the component tilt angle, determined by the conditions of the target construction area, respectively. and These are the lower limit and upper limit of component installation height, determined by the conditions of the target construction area, respectively. and These are the lower limit and upper limit of the array span, determined by the conditions of the target construction area, respectively; and These are the lower limit and upper limit of the lateral clearance of the components, determined by the conditions of the target construction area, respectively. A set of regional environmental and agricultural production constraint parameters for the target construction area.

[0029] Furthermore, the crop canopy light intensity and light uniformity in the multi-objective optimization model are calculated according to the following steps:

[0030] Calculate the first Crop canopy light intensity during each growth stage : In the formula, For the first Number of illumination sampling points on the canopy calculation plane during each reproductive period; For the first The number of time steps involved in the evaluation during the reproductive period; For the first One reproductive period; For the first The first crop canopy calculation plane at each growth stage Each sampling point at time... Illuminance;

[0031] based on Calculate the average canopy illuminance based on the number of light sampling points. ;

[0032] Calculate the first Standard deviation of canopy light rate during each reproductive period : ;

[0033] Calculate the first Light uniformity during the reproductive period : ;

[0034] Calculate the first Low light ratio during each reproductive period and longest continuous shade duration and satisfy the following constraints: In the formula, For the first Light requirement threshold for each reproductive period; For the first Low light ratio during the reproductive period; For the first The upper limit of the allowable low light ratio during the reproductive period; For the first The longest continuous shade duration during a reproductive period; For the first The maximum allowed continuous shade duration during a reproductive period.

[0035] Furthermore, the photovoltaic power generation per unit area, the crop canopy light intensity during the entire growth period, and the uniformity of light intensity during the entire growth period in the multi-objective optimization model are calculated according to the following steps:

[0036] Calculate photovoltaic power generation per unit area : In the formula, The area occupied by agricultural photovoltaic systems; The set of time steps involved in power generation evaluation over the entire reproductive period or the whole year; For photovoltaic module conversion efficiency; For photovoltaic system efficiency; This refers to the area of ​​a single photovoltaic module; This refers to the number of photovoltaic modules; For a moment with component tilt angle Array span lateral gap between components The amount of radiation received by the relevant components; For time step;

[0037] Calculate the amount of light received by the crop canopy throughout its entire growth period. : In the formula, The number of growth periods encompassed throughout the entire growth cycle of the target crop; For the first The evaluation weights for each reproductive period, and the evaluation weights satisfy the following: ;

[0038] Calculate the uniformity of light throughout the entire reproductive period : ;

[0039] Establish a multi-objective optimization model for the layout of the agricultural photovoltaic array: In the formula, Optimize variable vectors for layout Corresponding crop canopy light intensity throughout the entire growth period; Optimize variable vectors for layout The corresponding photovoltaic power generation per unit area; Optimize variable vectors for layout Corresponding uniformity of light throughout the entire reproductive period; The feasible layout parameter space is determined by the conditions of the target construction area, and the feasible layout parameter space Including each reproductive period The constraints include light threshold, low light ratio, and continuous shading duration, as well as constraints on the clearance of agricultural machinery, passage width, and installation boundaries of photovoltaic projects.

[0040] Furthermore, the establishment and validation of the response surface model based on the samples after partition error correction includes:

[0041] A response surface methodology is used to establish a response surface model between the layout variables and the objective function. This response surface model is a multinomial regression model containing linear, interaction, and quadratic terms, used to characterize the tilt angle of the photovoltaic module. Component installation height Array span and the lateral gaps between components With the amount of light received by the crop canopy at each growth stage Uniformity of light during each growth stage Low light ratio Continuous shade duration and photovoltaic power generation per unit area The mapping relationship between them;

[0042] The response surface model is determined according to the following formula: In the formula, For the first Each response surface output index takes the value of any one of the reproductive period indicators and the comprehensive indicator; For response indicator sequence number; and The tilt angles of the photovoltaic modules are respectively Component installation height Array span lateral gap between components Any variable in; and To optimize variable indexing; For the first The constant term regression coefficients of a response surface model; For the first Variables in a response surface model The regression coefficient of the first-order term; For the first Variables in a response surface model With variables The regression coefficients of the interaction term; For the first Variables in a response surface model The quadratic regression coefficient;

[0043] The output of the response surface model is first filtered by constraints such as light demand thresholds for each growth stage, low light ratio, and continuous shading duration, and then summarized into... , and Used for multi-objective optimization; after the response surface model is established, error verification is performed using simulation recalculation data or field measured data, when any growth stage... , , or When the prediction error exceeds the preset threshold, the corresponding variable combination is added to the sample set and the response surface model is refitted.

[0044] Furthermore, the non-dominated sorting genetic algorithm II is used to perform a global search in the feasible layout parameter space to obtain an initial Pareto solution set. This initial Pareto solution set is then used as the initial fish swarm for local updates in the artificial fish swarm algorithm to obtain a Pareto candidate layout solution set, including:

[0045] The objective function is uniformly transformed into minimizing the objective vector. : ;

[0046] The non-dominated sorting genetic algorithm II is used in the feasible layout parameter space. Perform a global search to obtain the initial Pareto solution set: In the formula, This is the initial Pareto solution set; This is a non-dominated sorting genetic algorithm II;

[0047] Will Each candidate layout, including component tilt angle, component mounting height, array span, and component lateral clearance, is used as the initial position of an artificial fish, and... , , and Mapped to respectively Normalized space, normalized variables are: In the formula, For the first Normalized values ​​of the optimization variables; To optimize the variable index, its corresponding variable is: , , or ; For the first The values ​​of each optimization variable in the original variable space; For the first The lower bound of each optimization variable; For the first The upper limit of the number of optimization variables;

[0048] In the normalized variable space, calculate the... Normalized view of optimization variables : In the formula, For the first The field scale coefficient of each optimization variable;

[0049] In the normalized variable space, calculate the... Normalization step size of each optimization variable : In the formula, For the first Step size scaling coefficients for each optimization variable;

[0050] In the original variable space, calculate the... Vision of an optimization variable : ;

[0051] In the original variable space, calculate the... Step size of each optimization variable : ;

[0052] The Pareto candidate layout solution set is updated according to the following formula: In the formula, This is the Pareto candidate layout solution set obtained after local updating using the artificial fish swarm algorithm; This is an artificial fish swarm algorithm; Indicates As the initial fish population, feasibility screening conditions were determined based on the light requirements of overwintering and summer-sown crops throughout their entire growth period, low light ratio, continuous shading duration, and agricultural machinery access constraints. Dominance relationships and crowding distance were used as movement selection criteria, and non-dominated ranking was used to update the population. The local neighborhood optimization process; wherein, when any candidate solution does not meet the light requirement threshold, the upper limit of low light ratio, or the upper limit of continuous shading duration during any critical growth period, it will not be included in the Pareto candidate layout solution set;

[0053] The local update includes foraging behavior, swarming behavior, tail-chasing behavior, and random behavior. In the foraging behavior, multiple candidate neighborhood solutions are randomly generated within the current field of view of the artificial fish, and solutions that satisfy these conditions are preferentially selected. The system identifies candidate solutions that dominate the current solution, considering crop light requirements at each growth stage and agricultural machinery passage constraints. When no dominant candidate solution exists, the solution with the lowest constraint violation and the largest congestion distance is selected. In clustering behavior, the center position of feasible non-dominated neighbors within the field of view is calculated, and movement is only performed if this center position does not cause constraint violation and the neighborhood density is below a preset congestion threshold. In tailgating behavior, the candidate solution with the largest congestion distance among feasible non-dominated neighbors within the field of view is selected as the tailgating target. New positions obtained from foraging, clustering, tailgating, or random behaviors are truncated and made feasible, eliminating candidate solutions that cause any growth stage light requirement threshold, agricultural machinery clearance, passage width, or variable out-of-bounds conditions to be met. Feasible candidate solutions are then merged with the current solution set, and the Pareto candidate layout solution set is updated after non-dominated sorting and congestion distance truncation.

[0054] Furthermore, the Pareto candidate layout solutions that satisfy the constraints are comprehensively evaluated using the entropy weight method and the approximation ideal solution ranking method to select the agricultural photovoltaic array layout scheme with the best overall performance, including:

[0055] Hard constraint filtering is performed based on the growth period constraint groups corresponding to overwintering crops, summer-sown crops, or crop rotation patterns. Candidate layouts that simultaneously meet the light requirement threshold, low light ratio upper limit, continuous shading duration upper limit, and agricultural machinery passage constraints for each growth period are included in the comprehensive evaluation.

[0056] Construct a system based on the total light intensity of the crop canopy throughout its entire growth period. Photovoltaic power generation per unit area and uniformity of light throughout the reproductive period The evaluation matrix consists of, where and As a benefit-oriented indicator, It is a cost-based indicator, and retains data for each reproductive stage. , , and As a constraint review indicator;

[0057] For benefit-related indicators, the standardized indicator values ​​are: For cost-related indicators, the standardized indicator value is: In the formula, For the first The Pareto candidate layout schemes are in the... The original indicator values ​​under each evaluation indicator; The Pareto candidate layout number; The evaluation indicator number; For use in The sequence number of the extreme value among the candidate layout schemes; The number of Pareto candidate layouts that are included in the comprehensive evaluation;

[0058] Calculate the first The candidate layout schemes are in the first The proportion of each evaluation indicator : ;

[0059] Calculate the first Information entropy of each evaluation indicator : Among them, when hour, Process as zero;

[0060] Calculate the first Objective weights of each evaluation indicator : In the formula, The number of evaluation indicators;

[0061] Calculate the distance from the candidate solution to the positive ideal solution. : In the formula, To evaluate the first in the matrix The candidate solution is in the... Standardized values ​​under each evaluation indicator; For the first The positive ideal solution for each evaluation index;

[0062] Calculate the distance from the candidate solution to the negative ideal solution. : In the formula, For the first The negative ideal solution of each evaluation index;

[0063] Calculate the relative similarity of each candidate layout scheme. : ;

[0064] Calculate the optimal agricultural photovoltaic array layout scheme based on overall performance. : In the formula, The optimal agricultural photovoltaic array layout scheme in terms of overall performance; For the first One Pareto candidate layout scheme.

[0065] Furthermore, the step of determining whether the verification error meets a preset threshold based on the light environment simulation verification data and field test verification data of the optimal agricultural photovoltaic array layout scheme includes: If the threshold is met, the optimal layout parameters are output; otherwise, supplementary sampling and reconstruction of the response surface model are performed.

[0066] Obtain the light environment simulation verification data and field test verification data corresponding to the agricultural photovoltaic array layout scheme with the optimal overall performance;

[0067] Recalculate canopy light intensity, light uniformity, low light ratio, continuous shading duration, and yield response according to the growth period;

[0068] When the verification error exceeds the preset threshold, supplementary sampling data will be added to the sample set and the response surface model will be reconstructed.

[0069] When the verification error meets the preset threshold, the output includes the tilt angle of the photovoltaic module, the installation height of the module, the array span, the lateral gap of the module, and the corresponding crop canopy light intensity, photovoltaic power generation per unit area, and light uniformity.

[0070] Based on the light environment simulation verification data and field trial verification data, the growth priority, light requirement threshold and layout adaptation strategy of overwintering crops and summer-sown crops are corrected.

[0071] Beneficial effects:

[0072] As can be seen from the above technical solutions, the present invention provides a multi-objective collaborative optimization method for agricultural photovoltaic arrays based on light energy coupling regulation, which has the following beneficial effects:

[0073] 1. This application acquires geographic and meteorological data, solar radiation data, photovoltaic module structural parameters, farmland spatial parameters, crop type, growth period data, canopy height parameters, light requirement threshold, and shading sensitivity coefficient for the target construction area. This data is then incorporated into a unified light environment simulation and parameter optimization process, ensuring that array layout design is no longer based solely on fixed empirical spacing or single module tilt angles. Since crop canopy height, solar altitude angle, and photovoltaic module projection range all vary over time, a unified data foundation allows candidate layouts to reflect the correlation between site conditions, meteorological conditions, crop conditions, and engineering conditions from the initial modeling stage, thereby reducing the problem of layout parameters deviating from agricultural production needs due to missing basic data.

[0074] 2. This application establishes a crop canopy calculation plane based on canopy height at different growth stages, and uses an unobstructed control area at the same time as a normalized benchmark to obtain illuminance, enabling comparison of illuminance data under different seasons, weather conditions, and solar altitude angles within the same evaluation scale. Since the illuminance inside agricultural photovoltaic arrays is affected by the component tilt angle, installation height, span, lateral gaps, and canopy position, the canopy planes for different growth stages can reflect the actual light exposure of crops during the seedling, vegetative growth, reproductive growth, and maturity stages. This approach helps avoid evaluation distortion caused by using a single canopy height or a single seasonal average, making the layout optimization results closer to the light requirements of crops throughout their entire growth period.

[0075] 3. This application statistically analyzes the measured errors in the area below the photovoltaic modules, the area between the photovoltaic modules, and the unshaded control area, and uses a partition error correction coefficient to correct the canopy illumination distribution data. The shading intensity, diffuse radiation ratio, and surface reflection conditions vary in different areas within an agricultural photovoltaic array. Using only a uniform error correction method may cause prediction biases in the area below and between the modules. Partition error correction makes the response surface training samples closer to the actual light environment. When any partition error exceeds a threshold, additional samples are added and the model is refitted, which reduces the prediction bias of the response surface model in local areas and provides a more reliable data foundation for subsequent multi-objective optimization.

[0076] 4. This application combines the global search capability of the Non-Dominated Sorting Genetic Algorithm II with the local neighborhood update capability of the Artificial Fish Swarm Algorithm. The Non-Dominated Sorting Genetic Algorithm II can obtain an initial Pareto solution set within the feasible layout parameter space, avoiding one-sided selection caused by single-objective optimization; the Artificial Fish Swarm Algorithm uses the initial Pareto solution set as the initial fish swarm, and introduces dominance relations, crowding distance, crop light requirement threshold, low light ratio, continuous shading duration, and agricultural machinery passage constraints in foraging, clustering, tail chasing, and random behaviors, ensuring that the local search does not deviate from the agricultural production boundary. This process can balance the distribution of the solution set and the ability to improve locally, thereby obtaining a candidate layout set that is more suitable for the common needs of crop growth and photovoltaic power generation.

[0077] 5. Before comprehensive evaluation, this application employs hard constraint filtering, incorporating only Pareto candidate layouts that meet the crop's light requirement threshold throughout its entire growth period, the upper limit of low light ratio, the upper limit of longest continuous shading duration, and the requirements for agricultural machinery passage into the evaluation using the entropy weight method and the approximation of the ideal solution ranking method. This process avoids selecting layout schemes that do not meet agricultural production conditions simply because a single indicator value is superior. Furthermore, crop canopy light intensity and photovoltaic power generation per unit area are used as benefit-type indicators, while light uniformity is used as a cost-type indicator. These are ranked using objective weights and the distance to the ideal solution. This allows for quantitative comparison of multi-objective performance while meeting the minimum requirements for agricultural production, thereby generating outputtable optimal layout parameters.

[0078] It should be understood that all combinations of the foregoing concepts and the additional concepts described in more detail below can be considered part of the inventive subject matter of this disclosure, provided that such concepts do not contradict each other.

[0079] The foregoing and other aspects, embodiments, and features of the teachings of the present invention will be more fully understood from the following description in conjunction with the accompanying drawings. Other additional aspects of the invention, such as features and / or beneficial effects of exemplary embodiments, will become apparent from the following description or may be learned through practice of specific embodiments according to the teachings of the present invention. Attached Figure Description

[0080] The accompanying drawings are not drawn to scale. In the drawings, each identical or nearly identical component shown in the various figures may be denoted by the same reference numeral. For clarity, not every component is labeled in each figure. Embodiments of various aspects of the invention will now be described by way of example and with reference to the accompanying drawings, wherein:

[0081] Figure 1 This is a flowchart illustrating a multi-objective collaborative optimization method for agricultural photovoltaic arrays based on light energy coupling regulation, as described in this application.

[0082] Figure 2 This is a flowchart of step S106 of a multi-objective collaborative optimization method for agricultural photovoltaic arrays based on light energy coupling regulation in an embodiment of this application.

[0083] Figure 3 This is a flowchart of step S110 of a multi-objective collaborative optimization method for agricultural photovoltaic arrays based on light energy coupling regulation in an embodiment of this application.

[0084] Figure 4 This is a flowchart of step S112 of a multi-objective collaborative optimization method for agricultural photovoltaic arrays based on light energy coupling regulation in an embodiment of this application.

[0085] Figure 5 This is a flowchart of step S114 of a multi-objective collaborative optimization method for agricultural photovoltaic arrays based on light energy coupling regulation in an embodiment of this application.

[0086] Figure 6 This is a flowchart of step S116 of a multi-objective collaborative optimization method for agricultural photovoltaic arrays based on light energy coupling regulation in an embodiment of this application.

[0087] Figure 7 This is a schematic diagram showing the distribution of Pareto candidate layout solutions in the parameter space for a multi-objective collaborative optimization method for agricultural photovoltaic arrays based on light energy coupling regulation, as described in an embodiment of this application.

[0088] Figure 8 This is a schematic diagram of an electronic device according to an embodiment of this application. Detailed Implementation

[0089] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Based on the described embodiments of the present invention, the technical or scientific terms used should have the ordinary meaning understood by those skilled in the art to which this invention pertains. Furthermore, the terms "first," "second," and similar words used in the patent application specification and claims of this invention do not indicate any order, quantity, or importance, but are only used to distinguish different components. Similarly, unless the context clearly indicates otherwise, the singular forms of "an," "a," or "the," etc., do not indicate a quantity limitation, but rather indicate the presence of at least one. Terms such as "comprising" or "including" indicate that the element or object preceding "comprising" encompasses the features, integrals, steps, operations, elements, and / or components listed following "comprising" or "including," and do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or sets thereof. Terms such as "upper," "lower," "left," and "right" are only used to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0090] Current agricultural photovoltaic array designs typically rely primarily on engineering experience, fixed spacing, tilt angle of individual modules, or single power generation targets, rarely considering the coupling relationship between canopy height variations throughout the crop's growth cycle, low light intensity, duration of continuous shading, light uniformity, agricultural machinery access space, and power generation per unit area. This approach can easily result in candidate layouts meeting requirements in a particular season or growth period, but exhibiting problems such as insufficient canopy light reception, prolonged localized shading, or decreased land power generation efficiency during other critical growth periods.

[0091] Therefore, embodiments of the present invention provide a multi-objective collaborative optimization method for agricultural photovoltaic arrays based on light energy coupling regulation, referring to... Figure 1 ,include:

[0092] Step S102: Obtain basic information data for optimizing the layout of agricultural photovoltaic arrays. The basic information data includes geographical and meteorological data of the target construction area, solar radiation data, photovoltaic module structural parameters, farmland spatial parameters, crop type, growth period data, canopy height parameters, light requirement threshold, and shading sensitivity coefficient.

[0093] Step S104: Establish a simulation model of the light environment of the agricultural photovoltaic system based on the basic information data, and establish a canopy calculation plane according to the canopy height at each growth stage. Using the unshaded control area at the same time as the normalization benchmark, obtain the canopy light distribution data, low light ratio, and continuous shading duration under each combination of layout parameters. Among them, the feasible layout parameter space is a set of layout parameters that simultaneously satisfy the following conditions: the canopy light intensity at each growth stage is not lower than the corresponding light requirement threshold; the low light ratio at each growth stage is not higher than the corresponding low light ratio upper limit; the longest continuous shading duration at each growth stage is not higher than the corresponding shading duration upper limit; and the requirements of agricultural machinery clearance, agricultural machinery passage width, and photovoltaic project installation boundary.

[0094] Step S106: Based on the measured errors of the area below the photovoltaic modules, the area between the photovoltaic modules, and the unshaded area, perform zoning error correction on the canopy illumination distribution data.

[0095] Step S108: Using the tilt angle of photovoltaic modules, the installation height of modules, the array span, and the lateral gap of modules as optimization variables, establish a multi-objective optimization model that includes crop canopy irradiance, photovoltaic power generation per unit area, and irradiance uniformity.

[0096] Step S110: Establish and validate the response surface model based on the samples after partition error correction.

[0097] Step S112: Use the non-dominated sorting genetic algorithm II to perform a global search in the feasible layout parameter space to obtain an initial Pareto solution set. Then, use the initial Pareto solution set as the initial fish swarm of the artificial fish swarm algorithm for local updates to obtain a Pareto candidate layout solution set.

[0098] Step S114: The Pareto candidate layout solution set that meets the light requirement threshold, low light ratio, continuous shading duration and agricultural machinery passage constraints of the entire crop growth period is comprehensively evaluated using the entropy weight method and the approximate ideal solution ranking method to select the agricultural photovoltaic array layout scheme with the best comprehensive performance.

[0099] Step S116: Based on the light environment simulation verification data and field test verification data of the agricultural photovoltaic array layout scheme with the best comprehensive performance, determine whether the verification error meets the preset threshold. If it does, output the optimal layout parameters; if it does not, supplement the sampling and reconstruct the response surface model.

[0100] First, basic data related to the target construction area, photovoltaic modules, farmland space, and crop growth stages are obtained. Then, a light environment simulation model based on the canopy height variation at each growth stage is established, and a time-synchronized normalization process is performed using an unshaded control area. Subsequently, the samples corrected for partitioning errors are used for response surface modeling. The tilt angle, installation height, array span, and lateral gap of the photovoltaic modules are used as optimization variables. Pareto candidate layout solutions are obtained through non-dominated sorting genetic algorithm II and artificial fish swarm algorithm. The layout scheme with better overall performance is selected through entropy weight method and approximation ideal solution sorting method. Finally, the output parameters are determined or the model is reconstructed by supplementing sampling based on the verification error.

[0101] Instead of designing separate optimization processes based on crop shading experience or photovoltaic power generation, this approach integrates crop growth period light intensity evaluation, photovoltaic power generation evaluation, light uniformity evaluation, and agricultural machinery access constraints into a closed-loop optimization process. It strings together canopy calculation planes for different growth periods, unshaded control normalization, zoning error correction, response surface proxy models, global search, local updates, and comprehensive evaluation into a parameter-based solution process adaptable to agricultural photovoltaics. This ensures that algorithm input, constraint screening, candidate solution updates, and final ranking are all controlled by agricultural production boundaries. Based on this, the natural conditions of the target construction area, photovoltaic module placement conditions, and crop growth conditions can be uniformly processed within the same process, reducing layout bias caused by single-objective optimization. By continuously incorporating crop light requirements, low light ratios, continuous shading duration, and agricultural machinery access constraints into candidate solution generation, candidate solution updates, and final evaluation, the final output layout parameters not only improve canopy light reception and light uniformity but also consider photovoltaic power generation capacity per unit land area, making it suitable for composite land use scenarios where agricultural production and photovoltaic power generation coexist.

[0102] In some embodiments, a simulation model of the light environment of the agricultural photovoltaic system is established based on basic information data, and a canopy calculation plane is established according to the canopy height at each growth stage, including:

[0103] Meteorological radiation data is cleaned and time-scaled, photovoltaic module structural parameters are converted, farmland space is 3D modeled or meshed, and crop canopy computational plane is discretely sampled.

[0104] A simulation model of the light environment of an agricultural photovoltaic system was established, including a three-dimensional structural model of photovoltaic modules, a farmland surface model, a crop canopy calculation plane, and solar radiation boundary conditions.

[0105] The entire growth period of the target crop is divided into several growth periods, and the first... Canopy height during each reproductive period or canopy height function and according to canopy height or canopy height function Establish a crop canopy computational plane ,in This refers to the reproductive period number. For evaluation time steps.

[0106] Calculation planes of each crop canopy The array is discretized into a grid based on planting row spacing, plant spacing, component projection boundary, and agricultural machinery passageway to extract the illuminance of each location within the agricultural photovoltaic array relative to the unshaded control area at the same time.

[0107] No. The first crop canopy calculation plane at each growth stage Each sampling point at time... Illuminance is determined according to the following formula: In the formula, For the first The first crop canopy calculation plane at each growth stage Each sampling point at time... Illuminance; The sampling point number; For the first agricultural photovoltaic array The reproductive period, the first Each sampling point, time The intensity of solar radiation or light intensity; This represents the solar radiation intensity or illumination intensity of the unobstructed control area at the same time. Simultaneous time-step normalization is used to eliminate the influence of different dates, weather conditions, and solar altitude angles on the absolute radiation intensity, enabling comparisons of canopy illumination under different combinations of layout parameters on a uniform scale.

[0108] By cleaning meteorological radiation data, standardizing the time scale, transforming photovoltaic module structural parameters, and performing 3D modeling or meshing of farmland space, a simulation model of the light environment of an agricultural photovoltaic system is established, including a 3D structural model of photovoltaic modules, a farmland surface model, a crop canopy calculation plane, and solar radiation boundary conditions. Canopy calculation planes of different heights are established based on canopy height or canopy height functions at different growth stages, and are then discretized into a mesh according to planting row spacing, plant spacing, module projection boundaries, and agricultural machinery passageways to extract the illuminance of each sampling point relative to the unshaded control area.

[0109] This application does not simplify the crop canopy to a fixed-height plane, but rather incorporates the characteristics of canopy height variation with the growth stage into the light environment calculation. The shading effect of agricultural photovoltaic arrays is closely related to the canopy position; the canopy heights differ at the seedling, jointing, flowering and fruiting, and maturity stages, and the relative positions of the photovoltaic module shadows to the canopy also differ. By establishing a canopy calculation plane according to the growth stage and normalizing it with an unshaded control area at the same time, the light differences under different seasons and canopy heights can be transformed into comparable illuminance indices, allowing subsequent optimization to avoid relying on a single absolute irradiance value.

[0110] This allows for the spatial, temporal, and growth-stage representation of the illumination status at various locations within the agricultural photovoltaic array. Since sampling points are arranged according to planting row spacing, plant spacing, component projection boundaries, and agricultural machinery access zones, the illumination evaluation results can reflect the actual planting space and engineering layout boundaries. Through synchronous time-synchronization, the interference of weather fluctuations and seasonal differences in radiation intensity on the evaluation results is reduced, thus providing a more consistent data basis for calculating low-light ratios, continuous shading durations, and total illumination throughout the growth period.

[0111] In some embodiments, the canopy illumination distribution data is corrected for regional errors based on the measured errors in the area below the photovoltaic modules, the area between the photovoltaic modules, and the unshaded area, referring to... Figure 2 ,include:

[0112] Step S1061: Obtain measured solar radiation data from solar radiation sensors deployed in the area below the photovoltaic modules, the area between the photovoltaic modules, and the unshaded control area.

[0113] Step S1062: Compare the measured solar radiation data with the output results of the agricultural photovoltaic system light environment simulation model, and calculate the relative error according to the area below the photovoltaic module, the area between the photovoltaic modules, and the unshaded area.

[0114] Step S1063: Establish partition error correction coefficients based on the relative errors of each partition, and correct the canopy illumination distribution data output by simulation using partition error correction coefficients.

[0115] Step S1064: When the relative error of any partition is greater than 8%, add the supplementary sample points of the corresponding partition to the sample set and regenerate the response surface training points until the corrected error meets the requirements of subsequent optimization calculation.

[0116] Solar radiation data was collected by solar radiation sensors deployed in the area below the photovoltaic modules, the area between the photovoltaic modules, and the unshaded control area. The measured data was compared with the output of the light environment simulation model, and the relative error of each zone was calculated. Then, a correction coefficient was established based on the relative error of each zone to correct the canopy illumination distribution data output by the simulation. When the relative error of any zone exceeds 8%, the corresponding zone's supplementary sample points were added to the sample set and the response surface training points were regenerated.

[0117] The error differences between different regions within the agricultural photovoltaic array are used as the basis for model correction. The area beneath the modules typically experiences strong shading, while the areas between modules are affected by both direct and diffuse radiation. Unshaded areas serve as a baseline. The sources of simulation error are not entirely the same for each region. Using a partitioned error correction coefficient avoids the accumulation of local biases caused by using a single error coefficient to correct the overall illumination data. When a threshold is exceeded, additional samples are added and training points are regenerated, enabling the response surface model to update for regions with larger errors.

[0118] The canopy illumination distribution data output by the simulation model can be calibrated using measured data, thereby improving the reliability of subsequent response surface modeling and multi-objective optimization results. The zonal correction method is beneficial for handling the spatial heterogeneity in agricultural photovoltaic systems, such as "long-term low illumination below modules" and "significant illumination variations in module spacing areas." Since error exceeding a threshold triggers sample replenishment and response surface training point updates, a feedback correction mechanism is formed, reducing computational bias caused by directly applying a fixed model to different plots or different meteorological conditions.

[0119] In some embodiments, a multi-objective optimization model is established using photovoltaic module tilt angle, module installation height, array span, and module lateral gap as optimization variables. This model includes crop canopy irradiance, photovoltaic power generation per unit area, and irradiance uniformity.

[0120] Set the tilt angle of the photovoltaic module to be The component installation height is The array span is The lateral gap between components is This forms the optimization variable vector: In the formula, Optimize the variable vector for the layout of agricultural photovoltaic arrays.

[0121] Among them, the optimization variable vector The following constraints must be satisfied: In the formula, and These are the lower limit and upper limit of the component tilt angle, determined by the conditions of the target construction area, respectively. and These are the lower limit and upper limit of component installation height, determined by the conditions of the target construction area, respectively. and These are the lower limit and upper limit of the array span, determined by the conditions of the target construction area, respectively; and These are the lower limit and upper limit of the lateral clearance of the components, determined by the conditions of the target construction area, respectively. A set of regional environmental and agricultural production constraint parameters for the target construction area.

[0122] The lower limit of the component tilt angle, the upper limit of the component tilt angle, the lower limit of the component installation height, the upper limit of the component installation height, the lower limit of the array span, the upper limit of the array span, the lower limit of the component lateral gap, and the upper limit of the component lateral gap are determined based on the latitude of the target area, solar radiation conditions, the light requirement threshold of crops at different growth stages, the canopy height variation curve, the allowable duration of continuous shading, the size of the photovoltaic module, the form of the support structure, the requirements for agricultural machinery passage, and the requirements for land use.

[0123] For overwintering crops, the key constraints for evaluation are the season with low solar altitude angle, the critical growth period from greening to grain filling, and the duration of long shadows; for summer-sown crops, the key constraints for evaluation are the season with high temperature and strong radiation, the critical growth period from seedling stage to flowering and fruiting, the stage of rapid canopy growth, and the tolerance range for moderate shading; when rotating crops on the same plot, the union of constraints for overwintering crops and summer-sown crops is used to form a shared feasible layout space, or seasonally adapted layout parameters are output separately.

[0124] The tilt angle of photovoltaic modules, installation height, array span, and lateral gap of modules constitute a vector of layout optimization variables, and the upper and lower limits of each variable are determined based on the conditions of the target construction area. These upper and lower limits are jointly determined by the latitude of the target area, solar radiation conditions, light requirements during the crop growth period, canopy height variation, allowable duration of continuous shading, photovoltaic module size, support structure, agricultural machinery access requirements, and land use requirements, ensuring that the optimization variables do not deviate from the boundaries of agricultural production and engineering installation.

[0125] The four selected optimization variables simultaneously influence both photovoltaic power generation capacity and crop canopy light reception. The module tilt angle affects the radiation received by the module and the length of the shadow projection; the installation height affects the distribution range of the shadow on the canopy plane; the array span affects the module density per unit land area and the crop's light-receiving space; and the lateral gap affects the light transmission path between modules and the power generation density. By unifying these four types of variables into a single vector and dynamically limiting their value range according to regional constraints, the coupled control relationship between agricultural photovoltaic array parameters can be demonstrated.

[0126] The optimization process can be carried out within clearly defined engineering feasibility boundaries, avoiding the acquisition of theoretically excellent but practically uninstallable or unsuitable parameters for agricultural machinery. By linking target area conditions with upper and lower limits of variables, the method can be adapted to agricultural photovoltaic systems of different latitudes, crop types, support structures, and plot sizes. This setup also facilitates the unified use of the same parameter space by subsequent response surface models, genetic algorithms, and artificial fish swarm algorithms, improving the consistency and executability of the computational process.

[0127] In some embodiments, crop canopy illuminance and light uniformity in the multi-objective optimization model are calculated according to the following steps:

[0128] Calculate the first Crop canopy light intensity during each growth stage : In the formula, For the first Number of illumination sampling points on the canopy calculation plane during each reproductive period; For the first The number of time steps involved in the evaluation during the reproductive period; For the first One reproductive period; For the first The first crop canopy calculation plane at each growth stage Each sampling point at time... Illumination rate.

[0129] based on Calculate the average canopy illuminance based on the number of light sampling points. .

[0130] Calculate the first Standard deviation of canopy light rate during each reproductive period : .

[0131] Calculate the first Light uniformity during the reproductive period : .

[0132] Calculate the first Low light ratio during each reproductive period and longest continuous shade duration and satisfy the following constraints: In the formula, For the first Light requirement threshold for each reproductive period; For the first Low light ratio during the reproductive period; For the first The upper limit of the allowable low light ratio during the reproductive period; For the first The longest continuous shade duration during a reproductive period; For the first The maximum allowed continuous shade duration during a reproductive period.

[0133] With the first The canopy computational plane of each growth period is used as the evaluation object. The illuminance of each sampling point and each time step during the growth period is statistically analyzed. The average illuminance reflects the overall light level of the canopy, and the standard deviation and coefficient of variation reflect the uniformity of light distribution in space and time. At the same time, the light demand threshold, the upper limit of low light ratio, and the upper limit of continuous shading duration are used as strong constraints for candidate layouts to enter into subsequent optimization.

[0134] The evaluation focuses not on a single moment or a single measuring point, but on statistically analyzing spatial sampling points and time steps throughout the growth cycle. While localized shading in agricultural photovoltaic systems may not significantly alter the overall average illumination, it can cause localized crops to experience prolonged periods of low light. Therefore, using only the average illumination is insufficient to determine whether a layout is suitable for agricultural production. Furthermore, by introducing the proportion of low illumination and the longest continuous shading duration, candidate layouts must meet three constraints: overall light exposure level, proportion of low-illuminated area, and temporal continuity. This allows for a more accurate description of the crop canopy's light exposure risk.

[0135] Based on the above design, layout schemes that meet the average light intensity requirement but suffer from prolonged local shading were eliminated during the optimization process. The impact of photovoltaic module placement on crop growth uniformity can be evaluated using the light uniformity index; the proportion of sampling points in the farmland with light requirements below the required level can be determined using the low light ratio index; and the longest continuous shading duration can identify the impact of sustained local shading on crop photosynthesis and growth rhythm. These three indices work together to make the subsequent Pareto solution more consistent with agricultural production constraints.

[0136] In some embodiments, the photovoltaic power generation per unit area, the crop canopy light intensity throughout the entire growth period, and the uniformity of light intensity throughout the entire growth period in the multi-objective optimization model are calculated according to the following steps:

[0137] Calculate photovoltaic power generation per unit area : In the formula, The area occupied by agricultural photovoltaic systems; The set of time steps involved in power generation evaluation over the entire reproductive period or the whole year; For photovoltaic module conversion efficiency; For photovoltaic system efficiency; This refers to the area of ​​a single photovoltaic module; This refers to the number of photovoltaic modules; For a moment with component tilt angle Array span lateral gap between components The amount of radiation received by the relevant components; For time step.

[0138] Calculate the amount of light received by the crop canopy throughout its entire growth period. : In the formula, The number of growth periods encompassed throughout the entire growth cycle of the target crop; For the first The evaluation weights for each reproductive period, and the evaluation weights satisfy the following: .

[0139] Calculate the uniformity of light throughout the entire reproductive period : .

[0140] Establish a multi-objective optimization model for agricultural photovoltaic array layout: In the formula, Optimize variable vectors for layout Corresponding crop canopy light intensity throughout the entire growth period; Optimize variable vectors for layout The corresponding photovoltaic power generation per unit area; Optimize variable vectors for layout Corresponding uniformity of light throughout the entire reproductive period; Let the feasible layout parameter space be determined by the conditions of the target construction area, and let the feasible layout parameter space be... Including each reproductive period The constraints include light threshold, low light ratio, and continuous shading duration, as well as constraints on the clearance of agricultural machinery, passage width, and installation boundaries of photovoltaic projects.

[0141] The conversion efficiency of the components, the system efficiency, the area of ​​the components, the number of components, the amount of radiation received by the components, and the time step are all included in the calculation of photovoltaic power generation per unit area. At the same time, the light intensity of the crop canopy and the uniformity of light intensity at each growth stage are re-aggregated according to the growth stage to form an evaluation index for the entire growth stage. A multi-objective optimization model is established to maximize the light intensity of the crop canopy, maximize the photovoltaic power generation per unit area, and minimize the uniformity of light intensity.

[0142] By unifying agricultural and photovoltaic (PV) evaluation indicators into a single optimization model, the agricultural side no longer considers only sunlight during a single growth period, but comprehensively considers the duration of sunlight, light sensitivity coefficient, and crop type through a weighted approach to the growth cycle. Similarly, the PV side does not only consider the power generation capacity of a single module, but uses PV power generation per unit area to evaluate land use efficiency. This is achieved by simultaneously setting... , and It can express the multi-objective conflict between crop light exposure, power generation, and light uniformity in agricultural photovoltaic systems.

[0143] The optimization results do not simply favor increasing array spacing to improve crop illumination, nor do they simply favor increasing component density to increase power generation. Instead, they seek coordination among multiple objectives within the feasible layout parameter space. By re-aggregating fertility options, the key fertility period can be made to have a corresponding impact on the final evaluation, adapting to the different light requirements of overwintering and summer-sown crops.

[0144] In some embodiments, a response surface model is established and validated based on samples corrected for partitioning errors, referring to... Figure 3 ,include:

[0145] Step S1101: Establish a response surface model between the layout variables and the objective function using the response surface methodology. The response surface model is a multinomial regression model containing linear, interaction, and quadratic terms, used to characterize the tilt angle of the photovoltaic module. Component installation height Array span and the lateral gaps between components With the amount of light received by the crop canopy at each growth stage Uniformity of light during each growth stage Low light ratio Continuous shade duration and photovoltaic power generation per unit area The mapping relationship between them.

[0146] The response surface model is determined according to the following formula: In the formula, For the first Each response surface output index takes the value of any one of the reproductive period indicators and the comprehensive indicator; For response indicator sequence number; and The tilt angles of the photovoltaic modules are respectively Component installation height Array span lateral gap between components Any variable in; and To optimize variable indexing; For the first The constant term regression coefficients of a response surface model; For the first Variables in a response surface model The regression coefficient of the first-order term; For the first Variables in a response surface model With variables The regression coefficients of the interaction term; For the first Variables in a response surface model The quadratic regression coefficient.

[0147] Step S1102: The output of the response surface model is first filtered by constraints such as light demand thresholds for each growth stage, low light ratio, and continuous shading duration, and then summarized as follows: , and Used for multi-objective optimization; after the response surface model is established, error verification is performed using simulation recalculation data or field measured data, when any growth stage... , , or When the prediction error exceeds the preset threshold, the corresponding variable combination is added to the sample set and the response surface model is refitted.

[0148] A multinomial regression model including linear, interaction, and quadratic terms is used to describe the mapping relationship between photovoltaic module tilt angle, module installation height, array span, and module lateral gap and crop canopy light intensity, light uniformity, low light ratio, continuous shading duration, and photovoltaic power generation per unit area. The response surface model output is first filtered by light requirement thresholds for each growth stage, low light ratio, and continuous shading duration constraints, and then summarized into a comprehensive index for multi-objective optimization.

[0149] Before global optimization, a response surface model corrected for measured errors is established to replace extensive repetitive light environment simulations and power generation simulations. Interactions exist among the layout parameters of agricultural photovoltaic arrays; for example, installation height and array span jointly affect shadow diffusion, while tilt angle and lateral clearance jointly affect the radiation received and light transmission paths of the modules. Employing a response surface model containing interaction terms and quadratic terms can more fully express the nonlinear relationships between variables than a single-factor linear model. Furthermore, error verification and sample supplementation improve model adaptability, increasing the efficiency of layout parameter search while ensuring computational feasibility. Performing complete light environment simulations and field verification for every candidate layout is computationally costly and unsuitable for iterative optimization. The response surface model can quickly predict multiple target indicators based on limited samples and determine whether additional samples are needed based on prediction errors. This mechanism ensures that the optimization process maintains reliance on physical simulations and measured results while meeting the computational needs of multi-objective algorithms for evaluating a large number of candidate solutions.

[0150] In some embodiments, a non-dominated sorting genetic algorithm II is used to perform a global search within the feasible layout parameter space to obtain an initial Pareto solution set. This initial Pareto solution set is then used as the initial fish swarm for local updates in an artificial fish swarm algorithm to obtain a Pareto candidate layout solution set. Figure 4 ,include:

[0151] Step S1121: Convert the objective function into minimizing the objective vector. : .

[0152] Step S1122: Use the non-dominated sorting genetic algorithm II in the feasible layout parameter space. Perform a global search to obtain the initial Pareto solution set: In the formula, This is the initial Pareto solution set; This is a non-dominated sorting genetic algorithm II.

[0153] Step S1123: ... Each candidate layout, including component tilt angle, component mounting height, array span, and component lateral clearance, is used as the initial position of an artificial fish, and... , , and Mapped to respectively Normalized space, normalized variables are: In the formula, For the first Normalized values ​​of the optimization variables; To optimize the variable index, its corresponding variable is: , , or ; For the first The values ​​of each optimization variable in the original variable space; For the first The lower bound of each optimization variable; For the first The upper limit of the number of optimization variables.

[0154] Step S1124: In the normalized variable space, calculate the first... Normalized view of optimization variables : In the formula, For the first The field scale coefficient of each optimization variable.

[0155] Step S1125: In the normalized variable space, calculate the first... Normalization step size of each optimization variable : In the formula, For the first The step size scaling coefficient of each optimization variable.

[0156] Step S1126: In the original variable space, calculate the first... Vision of an optimization variable : .

[0157] Step S1127: In the original variable space, calculate the first... Step size of each optimization variable : .

[0158] Step S1128: Update the Pareto candidate layout solution set according to the following formula: In the formula, This is the Pareto candidate layout solution set obtained after local updating using the artificial fish swarm algorithm; This is an artificial fish swarm algorithm; Indicates As the initial fish population, feasibility screening conditions were determined based on the light requirements of overwintering and summer-sown crops throughout their entire growth period, low light ratio, continuous shading duration, and agricultural machinery access constraints. Dominance relationships and crowding distance were used as movement selection criteria, and non-dominated ranking was used to update the population. The local neighborhood optimization process; wherein, when any candidate solution does not meet the light requirement threshold, the upper limit of low light ratio, or the upper limit of continuous shading duration during any critical growth period, it will not be included in the Pareto candidate layout solution set.

[0159] Local updates include foraging behavior, grouping behavior, tail-chasing behavior, and random behavior. In foraging behavior, multiple candidate neighborhood solutions are randomly generated within the current field of vision of the artificial fish, and solutions that satisfy these conditions are selected first. The system identifies candidate solutions that dominate the current solution, considering crop light requirements at each growth stage and agricultural machinery passage constraints. When no dominant candidate solution exists, the solution with the lowest constraint violation and the largest congestion distance is selected. In clustering behavior, the center position of feasible non-dominated neighbors within the field of view is calculated, and movement is only performed if this center position does not cause constraint violation and the neighborhood density is below a preset congestion threshold. In tailgating behavior, the candidate solution with the largest congestion distance among feasible non-dominated neighbors within the field of view is selected as the tailgating target. New positions obtained from foraging, clustering, tailgating, or random behaviors are truncated and made feasible, eliminating candidate solutions that cause any growth stage light requirement threshold, agricultural machinery clearance, passage width, or variable out-of-bounds conditions to be met. Feasible candidate solutions are then merged with the current solution set, and the Pareto candidate layout solution set is updated after non-dominated sorting and congestion distance truncation.

[0160] First, the maximization objective is transformed into the minimization objective vector. Then, the non-dominated sorting genetic algorithm II is used to perform a global search in the feasible layout parameter space to obtain the initial Pareto solution set. Subsequently, each candidate layout in the initial Pareto solution set is used as the initial position of the artificial fish, and it performs foraging, swarming, tail chasing and random behavior in the normalized variable space. The Pareto candidate layout solution set is updated through dominance relationship, crowding distance, constraint violation degree and boundary feasibility processing.

[0161] A combination of global search and local neighborhood update is used to solve for the layout parameters of agricultural photovoltaic arrays. The non-dominated sorting genetic algorithm II can form a widely distributed Pareto solution set within a large parameter space, but the accuracy of the local search may be affected by the population evolution process. The artificial fish swarm algorithm uses existing Pareto solutions as the initial fish swarm and can further explore the local neighborhood near feasible solutions. The crop light requirement threshold, low light ratio, continuous shading duration, and agricultural machinery passage constraints are embedded into the artificial fish movement selection criteria, so that the local update process is not driven solely by mathematical objectives.

[0162] This improves the distribution and local improvement capabilities of the candidate layout solution set. Normalization allows variables with different dimensions, such as inclination angle, height, span, and lateral clearance, to perform neighborhood searches on a unified scale; boundary truncation and feasibility processing prevent artificial fish movement from generating candidate solutions that are uninstallable or do not meet agricultural production constraints; non-dominated sorting and crowding distance truncation preserve candidate schemes with good multi-objective trade-offs, thus providing a high-quality Pareto candidate layout solution set for subsequent comprehensive evaluation.

[0163] In some embodiments, the Pareto candidate layout solutions that satisfy the constraints are comprehensively evaluated using the entropy weight method and the approximation ideal solution ranking method to select the agricultural photovoltaic array layout scheme with the best overall performance, referring to... Figure 5 ,include:

[0164] Step S1141: Perform hard constraint filtering based on the growth period constraint group corresponding to overwintering crops, summer-sown crops, or crop rotation patterns, and include candidate layouts that simultaneously meet the light requirement threshold, low light ratio upper limit, continuous shading duration upper limit, and agricultural machinery passage constraints of each growth period into the comprehensive evaluation.

[0165] Step S1142: Construct a model based on the total crop canopy light intensity throughout its growth period. Photovoltaic power generation per unit area and uniformity of light throughout the reproductive period The evaluation matrix consists of, where and As a benefit-oriented indicator, It is a cost-based indicator, and retains data for each reproductive stage. , , and As a constraint review indicator.

[0166] For benefit-type indicators, the standardized indicator values ​​are: For cost-related indicators, the standardized indicator value is: In the formula, For the first The Pareto candidate layout schemes are in the... The original indicator values ​​under each evaluation indicator; The Pareto candidate layout number; The evaluation indicator number; For use in The sequence number of the extreme value among the candidate layout schemes; This represents the number of Pareto candidate layout schemes that are included in the comprehensive evaluation.

[0167] Step S1143: Calculate the first... The candidate layout schemes are in the first The proportion of each evaluation indicator : .

[0168] Step S1144: Calculate the first... Information entropy of each evaluation indicator : Among them, when hour, Process as zero.

[0169] Step S1145: Calculate the first... Objective weights of each evaluation indicator : In the formula, The number of evaluation indicators.

[0170] Step S1146: Calculate the distance from the candidate solution to the positive ideal solution. : In the formula, To evaluate the first in the matrix The candidate solution is in the... Standardized values ​​under each evaluation indicator; For the first The positive ideal solution for each evaluation index.

[0171] Step S1147: Calculate the distance from the candidate solution to the negative ideal solution. : In the formula, For the first The negative ideal solution of each evaluation index.

[0172] Step S1148: Calculate the relative proximity of each candidate layout scheme. : .

[0173] Step S1149: Calculate the optimal agricultural photovoltaic array layout scheme based on overall performance. : In the formula, The optimal agricultural photovoltaic array layout scheme in terms of overall performance; For the first One Pareto candidate layout scheme.

[0174] First, hard constraint filtering is performed based on the growth period constraints corresponding to overwintering crops, summer-sown crops, or crop rotation patterns. Then, an evaluation matrix is ​​constructed, consisting of crop canopy irradiance throughout the entire growth period, photovoltaic power generation per unit area, and light uniformity throughout the entire growth period. Crop canopy irradiance and photovoltaic power generation per unit area are used as benefit indicators, while light uniformity is used as a cost indicator. The optimal layout scheme is then determined through standardization, information entropy, objective weights, distance between positive and negative ideal solutions, and relative proximity. The multi-objective Pareto solution set of agricultural photovoltaic array layout is further transformed into a sortable engineering decision result. The Pareto solution set typically contains multiple mutually indeterminate candidate schemes; the final output parameters cannot be directly determined solely by Pareto relations.

[0175] First, candidate solutions are filtered using hard constraints related to agricultural production. Then, the entropy weight method is used to determine objective weights based on the dispersion of indicators. Finally, the distance relationship between each scheme and the positive and negative ideal solutions is evaluated using the approximation-ideal-solution ranking method. This results in a comprehensive ranking without subjectively fixing a single weight. The final output layout scheme satisfies the agricultural production boundary and achieves a quantifiable comprehensive evaluation result in terms of crop light exposure, power generation per unit area, and light uniformity. Hard constraint filtering prevents schemes that do not meet the light requirements during the critical growth period from being selected due to their high power generation. The entropy weight method reduces the bias caused by human weighting. The approximation-ideal-solution ranking method transforms the superiority-inferiority relationship of multiple indicators into relative proximity, facilitating the direct selection of photovoltaic module tilt angle, module installation height, array span, and module lateral gap during the engineering implementation phase.

[0176] In some embodiments, based on the light environment simulation verification data and field trial verification data of the agricultural photovoltaic array layout scheme with optimal overall performance, it is determined whether the verification error meets a preset threshold. If it does, the optimal layout parameters are output; otherwise, supplementary sampling is performed and the response surface model is reconstructed, referring to... Figure 6 ,include:

[0177] Step S1161: Obtain the light environment simulation verification data and field test verification data corresponding to the agricultural photovoltaic array layout scheme with the best overall performance.

[0178] Step S1162: Recalculate canopy light intensity, light uniformity, low light ratio, duration of continuous shading, and yield response for each growth period. Do not use a single canopy height or a single seasonal average to replace the evaluation for the entire growth period.

[0179] Step S1163: When the verification error exceeds the preset threshold, supplementary sampling data is added to the sample set and the response surface model is reconstructed.

[0180] Step S1164: When the verification error meets the preset threshold, output the photovoltaic module tilt angle, module installation height, array span, module lateral gap, and the corresponding crop canopy light intensity, photovoltaic power generation per unit area, and light uniformity.

[0181] Step S1165: Based on the light environment simulation verification data and field trial verification data, correct the growth priority, light requirement threshold and layout adaptation strategy of overwintering crops and summer-sown crops respectively.

[0182] The following is a specific embodiment for illustration:

[0183] The target agricultural photovoltaic system is located in the middle and lower reaches of a certain river. The region has a subtropical monsoon climate with an average annual temperature of about 16℃ and an annual precipitation of about 1030mm.

[0184] The agricultural photovoltaic system is located in an area of ​​approximately 47 hectares. 2 The photovoltaic installation capacity is approximately 20MW, with an annual power generation of approximately 24 million kWh. In the initial photovoltaic array, the module installation height is approximately 2.5m, the module tilt angle is approximately 24°, the array span is approximately 7.0m, and the module vertical coverage is approximately 53.3%. The photovoltaic modules are polycrystalline silicon modules with a rated power of 265Wp, and the size of a single module is 1640mm×992mm×35mm.

[0185] Component tilt angle Set to three horizontal angles of 18°, 27°, and 36°, with an array span of... The component installation height is set to three levels: 8.0m, 9.0m, and 10.0m. The horizontal spacing between components is set at three levels: 2.5m, 3.0m, and 3.5m. The test was set to three levels: 0m, 0.2m, and 0.4m, and 29 initial simulation test schemes were generated using Box Behnken experimental design.

[0186] The population size of the non-dominated sorting genetic algorithm II was set to 300, the maximum number of generations was set to 250, the Pareto ratio was set to 0.70, the crossover probability was set to 0.80, and the mutation probability was set to 0.20. The artificial fish swarm algorithm had a fish population of 35, a variable population of 4, a maximum number of iterations of 100, a normalized field of view of 0.08, and a normalized step size of 0.05.

[0187] After updating the Pareto candidate layout solution set using non-dominated sorting and crowding distance truncation, the following results are obtained: Figure 7 The diagram shows the distribution of Pareto candidate layout solutions in parameter space. Each scatter point represents a candidate optimization scheme that satisfies the layout constraints of the agricultural photovoltaic array. The candidate schemes are jointly determined by layout variables such as the tilt angle of the photovoltaic modules, the installation height of the modules, the array span, and the lateral gap between the modules. The distribution of the scatter points in three-dimensional space reflects the trade-offs between different candidate layout schemes and objectives such as crop canopy irradiance, photovoltaic power generation per unit area, and irradiance uniformity. Figure 7 It can be seen that after the NSGA-II global search and AFSA local update, the candidate solutions are not randomly discretely distributed, but rather form a relatively concentrated Pareto solution set within a certain parameter range. This indicates that, under the constraints of crop light requirements throughout its entire growth period, low light ratio, continuous shading duration, agricultural machinery access, and photovoltaic project installation, there exists a feasible compromise optimization region for agricultural photovoltaic array layout. Candidate schemes within this region can achieve a relatively good balance between improving crop canopy light reception, maintaining a high photovoltaic power generation per unit area, and improving the spatial uniformity of light.

[0188] After optimization calculations, the optimal agricultural photovoltaic array layout scheme obtained in this embodiment is as follows: photovoltaic module tilt angle of 30.1°, array span of 9.1m, module installation height of 3.1m, and module lateral gap of 0.24m. Under this scheme, the crop canopy light irradiance is 80.1%, the photovoltaic power generation per unit area is 1.35 MW·h / ha, and the light uniformity is 18.3%. Compared with the initial layout scheme, the optimized scheme increases the crop canopy light irradiance by 4.6%, reduces the photovoltaic power generation per unit area by only 0.4 MW·h / ha, and reduces the light uniformity by 12.7%. These results indicate that through response surface proxy, Initializing AFSA, normalized neighborhood search, and constraint violation elimination can reduce invalid candidate solutions and maintain the distribution of Pareto solution sets, thereby improving crop canopy illumination conditions and spatial uniformity while maintaining high photovoltaic power generation performance.

[0189] After obtaining the optimal layout scheme with comprehensive performance, simulation verification data and field trial verification data of its light environment are acquired, and the canopy irradiance, light uniformity, low light ratio, continuous shading duration, and yield response are recalculated according to the growth stage. When the verification error exceeds the preset threshold, supplementary sampling data is added to the sample set and the response surface model is reconstructed; when the verification error meets the preset threshold, the photovoltaic module tilt angle, module installation height, array span, module lateral gap, and the corresponding crop canopy irradiance, photovoltaic power generation per unit area, and light uniformity are output.

[0190] The light environment simulation model and photovoltaic power generation simulation model were re-entered for verification. The recalculated results showed that the crop canopy irradiance was 81.5%, the photovoltaic power generation per unit area was 1.35 MW·h / ha, and the light uniformity was 18.9%. Compared with the prediction results of the optimized algorithm, the relative error of canopy irradiance was 1.7%, the relative error of photovoltaic power generation per unit area was 0%, and the relative error of light uniformity was 3.2%, all of which were below the 8% threshold, indicating that the method provided in this application has good prediction accuracy and stability.

[0191] Agricultural photovoltaic (PV) systems are affected by site conditions, weather fluctuations, crop growth, and component installation errors, and single response surface model predictions may deviate from actual light environments. By incorporating optimization result verification and model feedback updates as part of a closed-loop process, the optimal layout scheme is re-inputted into the light environment simulation model and verified through field trials. This allows the system to determine whether the model predictions meet error requirements; if not, additional sampling and reconstruction of the response surface model ensure that the next round of optimization is based on data closer to the actual scenario. Therefore, the engineering applicability of the final output layout parameters is improved. The verification step avoids limiting the optimization results to the response surface prediction level; instead, it verifies canopy irradiance, power generation per unit area, irradiance uniformity, low irradiance ratio, and continuous shading duration through recalculation and field measurements. The feedback update step allows the growth priority, light requirement threshold, and layout adaptation strategy for overwintering and summer-sown crops to be adjusted based on the verification results, thereby adapting to the optimization needs of agricultural PV arrays under different planting systems and climatic years.

[0192] The above verification results show that, through partition error correction, full growth period constraints, and NSGA-II-AFSA local update rules, agricultural photovoltaic shading partitions, crop growth period light requirements, continuous shading duration, and agricultural machinery operation boundaries are directly embedded into the candidate solution generation, movement, and update processes. Compared with methods that only evaluate the impact of layout on the light environment and power generation, this application can formulate layout adaptation strategies for the long shadow stage of overwintering crops with low solar altitude angles and the rapid growth stage of summer-sown crops with high canopies, respectively. This improves the light conditions of crop canopies within the agricultural photovoltaic system, reduces spatial differences in light intensity and response surface prediction errors, and to some extent mitigates the decline in crop yield caused by photovoltaic shading.

[0193] Another embodiment of the present invention provides a multi-objective collaborative optimization device for agricultural photovoltaic arrays based on light energy coupling regulation, comprising:

[0194] The data acquisition module is used to obtain basic information data for optimizing the layout of agricultural photovoltaic arrays. The basic information data includes geographical and meteorological data of the target construction area, solar radiation data, photovoltaic module structural parameters, farmland spatial parameters, crop type, growth period data, canopy height parameters, light requirement threshold, and shading sensitivity coefficient.

[0195] The first creation module is used to establish a simulation model of the light environment of the agricultural photovoltaic system based on basic information data, and to establish a canopy calculation plane according to the canopy height at each growth stage. Using the unshaded control area at the same time as the normalization benchmark, the module obtains the canopy light distribution data, low light ratio and continuous shading duration under each combination of layout parameters.

[0196] The correction module is used to correct the canopy illumination distribution data for regional errors based on the measured errors in the area below the photovoltaic modules, the area between the photovoltaic modules, and the unshaded area.

[0197] The second creation module is used to establish a multi-objective optimization model that includes crop canopy irradiance, photovoltaic power generation per unit area, and irradiance uniformity, using photovoltaic module tilt angle, module installation height, array span, and module lateral gap as optimization variables.

[0198] The third creation module is used to build and validate the response surface model based on the samples after partition error correction.

[0199] The update module is used to perform a global search in the feasible layout parameter space using the non-dominated sorting genetic algorithm II to obtain an initial Pareto solution set, and then use the initial Pareto solution set as the initial fish swarm of the artificial fish swarm algorithm for local updates to obtain a Pareto candidate layout solution set.

[0200] The evaluation module is used to comprehensively evaluate the Pareto candidate layout solutions that meet the constraints of crop light requirement threshold throughout the entire growth period, low light ratio, continuous shading duration, and agricultural machinery access, using the entropy weight method and the approximate ideal solution ranking method, and to select the agricultural photovoltaic array layout scheme with the best overall performance.

[0201] The judgment module is used to determine whether the verification error meets the preset threshold based on the light environment simulation verification data and field test verification data of the agricultural photovoltaic array layout scheme with the best comprehensive performance. If it meets the threshold, it outputs the optimal layout parameters; if it does not meet the threshold, it supplements sampling and reconstructs the response surface model.

[0202] It should be noted that although several units or sub-units of the device have been mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to embodiments of this application, the features and functions of two or more units described above can be embodied in one unit. Conversely, the features and functions of one unit described above can be further divided and embodied by multiple units.

[0203] Based on the same inventive concept as the above method embodiments, this application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it enables the electronic device to implement the control method described in the above embodiments.

[0204] In one embodiment, the electronic device may be a server, and in this embodiment, the structure of the electronic device may be as follows: Figure 8 As shown, it includes a memory, a communication module, and one or more processors.

[0205] Memory is used to store computer programs executed by the processor. Memory can be mainly divided into a program storage area and a data storage area. The program storage area can store the operating system and programs required to run instant messaging functions, etc.; the data storage area can store various instant messaging information and operation instruction sets, etc.

[0206] Memory can be volatile memory, such as random access memory (RAM); memory can also be non-volatile memory, such as read-only memory, flash memory, hard disk drive (HDD), or solid-state drive (SSD); or memory can be any other medium capable of carrying or storing a desired computer program having the form of instructions or data structures and accessible by a computer, but is not limited thereto. Memory can be a combination of the above-mentioned types of memory.

[0207] A processor may include one or more central processing units (CPUs) or digital processing units, etc. A processor is used to implement the aforementioned data processing methods when a computer program stored in memory is invoked.

[0208] The communication module is used to communicate with terminal devices and other servers.

[0209] This application embodiment does not limit the specific connection medium between the above-described memory, communication module, and processor. This application embodiment... Figure 8 The memory and processor are connected via a bus, and the bus is in... Figure 8 The connections between other components are illustrated with arrows and are for illustrative purposes only, not as limiting information. Buses can be categorized as address buses, data buses, control buses, etc. For ease of description, Figure 8 The text uses only one arrow to describe it, but does not indicate that there is only one bus or one type of bus.

[0210] Based on the same inventive concept as the above-described method embodiments, embodiments of the present invention also provide a computer-readable storage medium for storing a computer program. When the computer program is run on a computer, it enables an electronic device to implement the control methods described in the above embodiments. The computer-readable storage medium can be a readable signal medium or a readable storage medium. A readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof.

[0211] Based on the same inventive concept as the above-described method embodiments, embodiments of the present invention also provide a computer program product. The computer program product includes a computer program that, when run on an electronic device, causes the electronic device to perform the steps of the control methods described above according to various exemplary embodiments of this application. The program product may take the form of any combination of one or more readable media. These computer program commands can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to produce a machine, such that the commands executed by the processor of the computer or other programmable data processing device generate a process for implementing... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

Claims

1. A multi-objective collaborative optimization method for agricultural photovoltaic arrays based on light energy coupling regulation, characterized in that, include: Acquire basic information data for optimizing the layout of agricultural photovoltaic arrays. The basic information data includes geographical and meteorological data of the target construction area, solar radiation data, photovoltaic module structural parameters, farmland spatial parameters, crop type, growth period data, canopy height parameters, light requirement threshold, and shading sensitivity coefficient. Based on the aforementioned basic information data, an agricultural photovoltaic system light environment simulation model was established, and a canopy calculation plane was established according to the canopy height at each growth stage. The unshaded control area at the same time was used as the normalization benchmark to obtain canopy light distribution data, low light ratio, and continuous shading duration under each combination of layout parameters. Based on the measured errors in the area below the photovoltaic modules, the area between the photovoltaic modules, and the unshaded area, the canopy illumination distribution data is corrected for partitioning errors. Using the tilt angle of photovoltaic modules, the installation height of modules, the array span, and the lateral gap of modules as optimization variables, a multi-objective optimization model is established, which includes crop canopy irradiance, photovoltaic power generation per unit area, and irradiance uniformity. A response surface model was established and validated based on the samples after partitioning error correction. The non-dominated sorting genetic algorithm II is used to perform a global search in the feasible layout parameter space to obtain an initial Pareto solution set. The initial Pareto solution set is then used as the initial fish swarm of the artificial fish swarm algorithm for local updates to obtain a Pareto candidate layout solution set. The Pareto candidate layout solution set that meets the light requirement threshold, low light ratio, continuous shading duration and agricultural machinery passage constraints of the entire crop growth period is comprehensively evaluated by the entropy weight method and the approximate ideal solution ranking method, and the agricultural photovoltaic array layout scheme with the best comprehensive performance is selected. Based on the light environment simulation verification data and field test verification data of the optimal agricultural photovoltaic array layout scheme, determine whether the verification error meets the preset threshold. If it does, output the optimal layout parameters; otherwise, supplement the sampling and reconstruct the response surface model.

2. The multi-objective collaborative optimization method for agricultural photovoltaic arrays based on light energy coupling regulation according to claim 1, characterized in that, The step of establishing a simulation model of the light environment of an agricultural photovoltaic system based on the aforementioned basic information data, and establishing a canopy calculation plane according to the canopy height at each growth stage, includes: Meteorological radiation data is cleaned and time-scaled, photovoltaic module structural parameters are converted, farmland space is modeled in three dimensions or meshed, and crop canopy calculation plane is discretized. Establish a simulation model of the light environment of an agricultural photovoltaic system, including a three-dimensional structural model of photovoltaic modules, a farmland surface model, a crop canopy calculation plane, and solar radiation boundary conditions; The entire growth period of the target crop is divided into several growth periods, and the first... Canopy height during each reproductive period or canopy height function and according to the canopy height or canopy height function Establish a crop canopy computational plane ,in This refers to the reproductive period number. For evaluation time steps; Calculation planes of each crop canopy The grid is discretized according to the planting row spacing, plant spacing, component projection boundary and agricultural machinery passage belt, and used to extract the illuminance of each position inside the agricultural photovoltaic array relative to the unshaded control area at the same time. No. The first crop canopy calculation plane at each growth stage Each sampling point at time... Illuminance is determined according to the following formula: In the formula, For the first The first crop canopy calculation plane at each growth stage Each sampling point at time... Illuminance; The sampling point number; For the first agricultural photovoltaic array The reproductive period, the first Each sampling point, time The intensity of solar radiation or light intensity; The solar radiation intensity or light intensity of the unobstructed control area at the same time.

3. The multi-objective collaborative optimization method for agricultural photovoltaic arrays based on light energy coupling regulation according to claim 2, characterized in that, The step of correcting the canopy illumination distribution data for regional errors based on the measured errors in the area below the photovoltaic modules, the area between the photovoltaic modules, and the unshaded area includes: Acquire measured solar radiation data from solar radiation sensors deployed in the area below the photovoltaic modules, the area between the photovoltaic modules, and the unshaded control area; The measured solar radiation data are compared with the output results of the agricultural photovoltaic system light environment simulation model, and the relative error is calculated according to the area below the photovoltaic module, the area between the photovoltaic modules, and the unshaded area. Based on the relative error of each partition, a partition error correction coefficient is established, and the canopy illumination distribution data output by the simulation is corrected using the partition error correction coefficient. When the relative error of any partition is greater than 8%, the supplementary sample points of the corresponding partition are added to the sample set and the response surface training points are regenerated until the corrected error meets the requirements of subsequent optimization calculations.

4. The multi-objective collaborative optimization method for agricultural photovoltaic arrays based on light energy coupling regulation according to claim 3, characterized in that, The aforementioned multi-objective optimization model, using photovoltaic module tilt angle, module installation height, array span, and module lateral gap as optimization variables, includes crop canopy irradiance, photovoltaic power generation per unit area, and irradiance uniformity. Set the tilt angle of the photovoltaic module to be The component installation height is The array span is The lateral gap between components is This forms the optimization variable vector: In the formula, Optimize the variable vector for agricultural photovoltaic array layout; The optimized variable vector The following constraints must be satisfied: In the formula, and These are the lower limit and upper limit of the component tilt angle, determined by the conditions of the target construction area, respectively. and These are the lower limit and upper limit of component installation height, determined by the conditions of the target construction area, respectively. and These are the lower limit and upper limit of the array span, determined by the conditions of the target construction area, respectively; and These are the lower limit and upper limit of the lateral clearance of the components, determined by the conditions of the target construction area, respectively. A set of regional environmental and agricultural production constraint parameters for the target construction area.

5. The multi-objective collaborative optimization method for agricultural photovoltaic arrays based on light energy coupling regulation according to claim 4, characterized in that, The crop canopy light intensity and light uniformity in the multi-objective optimization model are calculated according to the following steps: Calculate the first Crop canopy light intensity during each growth stage : In the formula, For the first Number of illumination sampling points on the canopy calculation plane during each reproductive period; For the first The number of time steps involved in the evaluation during the reproductive period; For the first One reproductive period; For the first The first crop canopy calculation plane at each growth stage Each sampling point at time... Illuminance; based on Calculate the average canopy illuminance based on the number of light sampling points. ; Calculate the first Standard deviation of canopy light rate during each reproductive period : ; Calculate the first Light uniformity during the reproductive period : ; Calculate the first Low light ratio during each reproductive period and longest continuous shade duration and satisfy the following constraints: In the formula, For the first Light requirement threshold for each reproductive period; For the first Low light ratio during the reproductive period; For the first The upper limit of the allowable low light ratio during the reproductive period; For the first The longest continuous shade duration during a reproductive period; For the first The maximum allowed continuous shade duration during a reproductive period.

6. The multi-objective collaborative optimization method for agricultural photovoltaic arrays based on light energy coupling regulation according to claim 5, characterized in that, The photovoltaic power generation per unit area, crop canopy light intensity during the entire growth period, and light uniformity during the entire growth period in the multi-objective optimization model are calculated according to the following steps: Calculate photovoltaic power generation per unit area : In the formula, The area occupied by agricultural photovoltaic systems; The set of time steps involved in power generation evaluation over the entire reproductive period or the whole year; For photovoltaic module conversion efficiency; For photovoltaic system efficiency; This refers to the area of ​​a single photovoltaic module; This refers to the number of photovoltaic modules; For a moment with component tilt angle Array span lateral gap between components The amount of radiation received by the relevant components; For time step; Calculate the amount of light received by the crop canopy throughout its entire growth period. : In the formula, The number of growth periods encompassed throughout the entire growth cycle of the target crop; For the first The evaluation weights for each reproductive period, and the evaluation weights satisfy the following: ; Calculate the uniformity of light throughout the entire reproductive period : ; Establish a multi-objective optimization model for the layout of the agricultural photovoltaic array: In the formula, Optimize variable vectors for layout Corresponding crop canopy light intensity throughout the entire growth period; Optimize variable vectors for layout The corresponding photovoltaic power generation per unit area; Optimize variable vectors for layout Corresponding uniformity of light throughout the entire reproductive period; The feasible layout parameter space is determined by the conditions of the target construction area, and the feasible layout parameter space Including each reproductive period The constraints include light threshold, low light ratio, and continuous shading duration, as well as constraints on the clearance of agricultural machinery, passage width, and installation boundaries of photovoltaic projects.

7. The multi-objective collaborative optimization method for agricultural photovoltaic arrays based on light energy coupling regulation according to claim 6, characterized in that, The process of establishing and validating the response surface model based on the samples after partition error correction includes: A response surface methodology is used to establish a response surface model between the layout variables and the objective function. This response surface model is a multinomial regression model containing linear, interaction, and quadratic terms, used to characterize the tilt angle of the photovoltaic module. Component installation height Array span and the lateral gaps between components With the amount of light received by the crop canopy at each growth stage Uniformity of light during each growth stage Low light ratio Continuous shade duration and photovoltaic power generation per unit area The mapping relationship between them; The response surface model is determined according to the following formula: In the formula, For the first Each response surface output index takes the value of any one of the reproductive period indicators and the comprehensive indicator; For response indicator sequence number; and The tilt angles of the photovoltaic modules are respectively Component installation height Array span lateral gap between components Any variable in; and To optimize variable indexing; For the first The constant term regression coefficients of a response surface model; For the first Variables in a response surface model The regression coefficient of the first-order term; For the first Variables in a response surface model With variables The regression coefficients of the interaction term; For the first Variables in a response surface model The quadratic regression coefficient; The output of the response surface model is first filtered by constraints such as light demand thresholds for each growth stage, low light ratio, and continuous shading duration, and then summarized into... , and Used for multi-objective optimization; after the response surface model is established, error verification is performed using simulation recalculation data or field measured data, when any growth stage... , , or When the prediction error exceeds the preset threshold, the corresponding variable combination is added to the sample set and the response surface model is refitted.

8. The multi-objective collaborative optimization method for agricultural photovoltaic arrays based on light energy coupling regulation according to claim 7, characterized in that, The non-dominated sorting genetic algorithm II is used to perform a global search in the feasible layout parameter space to obtain an initial Pareto solution set. This initial Pareto solution set is then used as the initial fish swarm for local updates in the artificial fish swarm algorithm to obtain a Pareto candidate layout solution set, including: The objective function is uniformly transformed into minimizing the objective vector. : ; The non-dominated sorting genetic algorithm II is used in the feasible layout parameter space. Perform a global search to obtain the initial Pareto solution set: In the formula, This is the initial Pareto solution set; This is a non-dominated sorting genetic algorithm II; Will Each candidate layout, including component tilt angle, component mounting height, array span, and component lateral clearance, is used as the initial position of an artificial fish, and... , , and Mapped to respectively Normalized space, normalized variables are: In the formula, For the first Normalized values ​​of the optimization variables; To optimize the variable index, its corresponding variable is: , , or ; For the first The values ​​of each optimization variable in the original variable space; For the first The lower bound of each optimization variable; For the first The upper limit of the number of optimization variables; In the normalized variable space, calculate the... Normalized view of optimization variables : In the formula, For the first The field scale coefficient of each optimization variable; In the normalized variable space, calculate the... Normalization step size of each optimization variable : In the formula, For the first Step size scaling coefficients for each optimization variable; In the original variable space, calculate the... Vision of an optimization variable : ; In the original variable space, calculate the... Step size of each optimization variable : ; The Pareto candidate layout solution set is updated according to the following formula: In the formula, This is the Pareto candidate layout solution set obtained after local updating using the artificial fish swarm algorithm; This is an artificial fish swarm algorithm; Indicates As the initial fish population, feasibility screening conditions were determined based on the light requirements of overwintering and summer-sown crops throughout their entire growth period, low light ratio, continuous shading duration, and agricultural machinery access constraints. Dominance relationships and crowding distance were used as movement selection criteria, and non-dominated ranking was used to update the population. The local neighborhood optimization process; wherein, when any candidate solution does not meet the light requirement threshold, the upper limit of low light ratio, or the upper limit of continuous shading duration during any critical growth period, it will not be included in the Pareto candidate layout solution set; The local update includes foraging behavior, swarming behavior, tail-chasing behavior, and random behavior. In the foraging behavior, multiple candidate neighborhood solutions are randomly generated within the current field of view of the artificial fish, and solutions that satisfy these conditions are preferentially selected. The system identifies candidate solutions that dominate the current solution, considering crop light requirements at each growth stage and agricultural machinery passage constraints. When no dominant candidate solution exists, the solution with the lowest constraint violation and the largest congestion distance is selected. In clustering behavior, the center position of feasible non-dominated neighbors within the field of view is calculated, and movement is only performed if this center position does not cause constraint violation and the neighborhood density is below a preset congestion threshold. In tailgating behavior, the candidate solution with the largest congestion distance among feasible non-dominated neighbors within the field of view is selected as the tailgating target. New positions obtained from foraging, clustering, tailgating, or random behaviors are truncated and made feasible, eliminating candidate solutions that cause any growth stage light requirement threshold, agricultural machinery clearance, passage width, or variable out-of-bounds conditions to be met. Feasible candidate solutions are then merged with the current solution set, and the Pareto candidate layout solution set is updated after non-dominated sorting and congestion distance truncation.

9. A multi-objective collaborative optimization method for agricultural photovoltaic arrays based on light energy coupling regulation according to claim 8, characterized in that, The process of comprehensively evaluating the Pareto candidate layout solutions that satisfy the constraints using the entropy weight method and the approximation ideal solution ranking method to select the agricultural photovoltaic array layout scheme with the best overall performance includes: Hard constraint filtering is performed based on the growth period constraint groups corresponding to overwintering crops, summer-sown crops, or crop rotation patterns. Candidate layouts that simultaneously meet the light requirement threshold, low light ratio upper limit, continuous shading duration upper limit, and agricultural machinery passage constraints for each growth period are included in the comprehensive evaluation. Construct a system based on the total light intensity of the crop canopy throughout its entire growth period. Photovoltaic power generation per unit area and uniformity of light throughout the reproductive period The evaluation matrix consists of, where and As a benefit-oriented indicator, It is a cost-based indicator, and retains data for each reproductive stage. , , and As a constraint review indicator; For benefit-related indicators, the standardized indicator values ​​are: For cost-related indicators, the standardized indicator value is: In the formula, For the first The Pareto candidate layout schemes are in the... The original indicator values ​​under each evaluation indicator; The Pareto candidate layout number; The evaluation indicator number; For use in The sequence number of the extreme value among the candidate layout schemes; The number of Pareto candidate layouts that are included in the comprehensive evaluation; Calculate the first The candidate layout schemes are in the first The proportion of each evaluation indicator : ; Calculate the first Information entropy of each evaluation indicator : Among them, when hour, Process as zero; Calculate the first Objective weights of each evaluation indicator : In the formula, The number of evaluation indicators; Calculate the distance from the candidate solution to the positive ideal solution. : In the formula, To evaluate the first in the matrix The candidate solution is in the... Standardized values ​​under each evaluation indicator; For the first The positive ideal solution for each evaluation index; Calculate the distance from the candidate solution to the negative ideal solution. : In the formula, For the first The negative ideal solution of each evaluation index; Calculate the relative similarity of each candidate layout scheme. : ; Calculate the optimal agricultural photovoltaic array layout scheme based on overall performance. : In the formula, The optimal agricultural photovoltaic array layout scheme in terms of overall performance; For the first One Pareto candidate layout scheme.

10. A multi-objective collaborative optimization method for agricultural photovoltaic arrays based on light energy coupling regulation according to any one of claims 1 to 9, characterized in that, The process involves using simulation data of the light environment and field test data of the optimal agricultural photovoltaic array layout scheme to determine whether the verification error meets a preset threshold. If it does, the optimal layout parameters are output; otherwise, additional sampling and reconstruction of the response surface model are performed. Obtain the light environment simulation verification data and field test verification data corresponding to the agricultural photovoltaic array layout scheme with the optimal overall performance; Recalculate canopy light intensity, light uniformity, low light ratio, continuous shading duration, and yield response according to the growth period; When the verification error exceeds the preset threshold, supplementary sampling data will be added to the sample set and the response surface model will be reconstructed. When the verification error meets the preset threshold, the output includes the tilt angle of the photovoltaic module, the installation height of the module, the array span, the lateral gap of the module, and the corresponding crop canopy light intensity, photovoltaic power generation per unit area, and light uniformity. Based on the light environment simulation verification data and field trial verification data, the growth priority, light requirement threshold and layout adaptation strategy of overwintering crops and summer-sown crops are corrected.

Images