A photovoltaic power station local population intelligent configuration repair method and system

Abstract

This application belongs to the interdisciplinary field of ecological restoration and artificial intelligence, specifically involving a method and system for intelligent configuration and restoration of native plant communities in photovoltaic power plants. It aims to solve problems such as low vegetation survival rates, severe soil erosion, and high operation and maintenance costs caused by traditional uniform sowing. The method includes: constructing a micro-habitat zoning model based on the height and tilt angle of photovoltaic modules, dividing the area into shaded, drip, and full-sunlight zones; quantifying the environmental pressure characteristics of each zone, such as photosynthetically active radiation, soil moisture, and erosion force; screening shade-tolerant, drought-tolerant, or erosion-resistant species from a local native plant bank; constructing a multi-objective function with survival rate, soil and water conservation, and functional diversity as optimization objectives, and using a non-dominated sorting genetic algorithm to solve for the optimal seed ratio in each zone; and implementing differentiated and precise sowing through positioning equipment. This application improves vegetation survival rate and ecological adaptability, effectively prevents soil erosion, reduces operation and maintenance costs, and constructs a sustainable native plant community system.

Description

Technical Field

[0001] This application belongs to the field of interdisciplinary technology of ecological restoration and artificial intelligence, specifically involving a method and system for intelligent configuration and restoration of native species populations in photovoltaic power stations. Background Technology

[0002] With the rapid development of the renewable energy industry, the large-scale deployment of photovoltaic (PV) power plants has become a key path to achieving the "dual carbon" goal. However, the construction of PV infrastructure has significantly altered the energy and water distribution patterns of the original surface ecosystem. Under typical ground-mounted PV arrays, the shading effect of the modules on solar radiation leads to high spatial heterogeneity in surface illumination. Simultaneously, their water collection effect causes rainfall to concentrate and drip along the edges of the modules, forming localized high-intensity erosion zones, while the area directly beneath the modules remains arid due to the difficulty in receiving rainwater. This artificially induced microhabitat differentiation, induced by engineering structures, poses a severe challenge to vegetation restoration.

[0003] The core of ecological restoration of photovoltaic power plants lies in the precise adaptation of vegetation configuration to local environmental conditions. Current restoration practices generally adopt a uniform sowing model across the entire site, ignoring the significant differences in light intensity, soil moisture, and hydrodynamic disturbance between shaded areas, dripping areas, and fully illuminated areas. This extensive strategy leads to stunted growth and even death of plants in shaded areas due to insufficient photosynthetically active radiation; exacerbated soil erosion in dripping areas due to a lack of erosion-resistant species cover; and resource waste and low survival rates in fully illuminated areas due to the failure to specifically select drought-resistant species. The superposition of multiple ecological stresses not only weakens the stability of the vegetation system but may also trigger engineering safety risks such as foundation settlement.

[0004] Therefore, there is an urgent need for a method for intelligent configuration and restoration of native species that can respond to the spatial heterogeneity of microhabitats in photovoltaic power plants, so as to achieve synergistic improvement of ecological function and engineering safety. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for intelligent configuration and repair of native species in photovoltaic power plants, which can effectively solve the problems mentioned in the background art.

[0006] The first aspect is the intelligent configuration and restoration method for native species populations in photovoltaic power plants disclosed in this application, which includes the following specific steps: Obtain the geometric parameters of the photovoltaic array, which include at least the height above the ground and the installation tilt angle; Based on the ground clearance and installation tilt angle, combined with the solar altitude angle, the boundary of the shadow area formed by the photovoltaic module is calculated; The maximum splash radius of raindrops under rainfall conditions was determined based on hydrodynamic sputtering experiments, and the boundary of the dripping area located at the lower edge of the component was delineated based on the maximum splash radius. The area within the station excluding the shaded area and the dripping area is divided into a fully lit area, and a microhabitat partitioning model including the shaded area, the dripping area and the fully lit area is constructed. Environmental pressure characteristic data of the shaded area, dripping area and full-light area were collected and quantified respectively. The environmental pressure characteristic data included at least photosynthetically active radiation, soil moisture and scouring force of dripping area. From the local native plant bank, suitable candidate species were selected based on environmental stress characteristic data of each zone; Based on the selected candidate species, the optimal seed ratio for each partition is obtained. Based on the optimal seed ratio and the microhabitat zoning model, differentiated precision sowing is carried out in different zones.

[0007] Furthermore, the environmental pressure characteristic data are quantified by constructing a three-dimensional environmental pressure characteristic vector, which includes cumulative photosynthetically active radiation, soil moisture availability index, and representative scour force.

[0008] Furthermore, suitable candidate species were screened from the local native plant bank, including: For the shaded area, shade tolerance screening conditions are set, including: under shading conditions where the photosynthetically active radiation value is 20% of that in the full-light area, the net photosynthetic rate maintenance rate is not less than 60%, and the ratio of chlorophyll b to chlorophyll a is greater than 1.2. For the dripping area, erosion resistance screening conditions are set, including: the root density is not lower than a preset first threshold, and the root tensile strength is not lower than a preset second threshold; For the full-sunlight zone, drought resistance screening conditions were set, including: when the soil moisture availability index is not greater than 0.25, the plant stomatal conductance is greater than 0.1 moles per square meter per second and the relative water content of the leaves is greater than 70%.

[0009] Furthermore, the optimal seed ratio for each partition is obtained by solving the problem, specifically: A multi-objective function is constructed with the optimization objectives of maximizing the weighted sum of expected vegetation survival rates in each zone, minimizing soil erosion across the entire station, and maximizing the functional diversity index of plant communities. Set optimization constraints that include a maximum total seeding amount, minimum coverage of a single species, and species compatibility. The multi-objective function is solved using a non-dominated sorting genetic algorithm to obtain a Pareto optimal solution set, and the solution with the highest comprehensive evaluation value is selected from the Pareto optimal solution set as the optimal seed ratio.

[0010] Furthermore, the total soil and water loss of the entire station is minimized by coupling the root system soil-fixing factor with the modified general soil loss equation RUSLE. The vegetation cover management factor is calculated by empirical formula based on the weighted average of the root density of species in each zone.

[0011] Furthermore, implementing differentiated precision seeding includes: The optimal seed ratio is associated with the digital vector map of the microhabitat zoning model to generate a work file containing zoning identifiers, spatial ranges, and corresponding seeding quantity lists; The seeding equipment, equipped with a positioning module and a multi-channel independent sowing device, automatically switches the corresponding seed storage bin or adjusts the sowing parameters based on real-time positioning information when the equipment enters different zones, according to the operation document, to ensure that the sowing amount per unit area is consistent with the optimal seed ratio.

[0012] Furthermore, after implementing differentiated precision seeding, it also includes post-implementation monitoring and feedback adjustment steps: At preset time points, the survival rate of vegetation in each zone was monitored to obtain the measured survival rate. The measured survival rate is compared with the expected survival rate predicted when solving the optimal seed ratio. If the deviation exceeds the preset threshold, the input data used for species screening or optimal ratio solving is corrected, and the screening and solving steps are re-executed to generate an updated seed formula.

[0013] Furthermore, monitoring vegetation survival rate in each zone includes: obtaining measured data through ground quadrat surveys, calculating the Normalized Difference Vegetation Index (NDVI) by combining UAV multispectral remote sensing images, using machine learning models to invert and obtain the vegetation cover distribution map of the entire field, and extracting the measured cover and survival rate of each zone from it.

[0014] Furthermore, the maximum splash radius was determined based on hydrodynamic splashing experiments. Specifically, under laboratory conditions simulating preset extreme rainfall intensity, high-speed camera technology was used to determine the maximum splash distance of rainwater droplets on surface soil particles, and this distance was used as the basis for delineating the dripping zone.

[0015] On the other hand, the intelligent configuration and restoration system for native species populations at photovoltaic power plants disclosed in this application includes: The data acquisition and partitioning module is used to obtain the geometric parameters of the photovoltaic array and, based on the geometric parameters, solar altitude angle, and maximum sputtering radius determined by hydrodynamic sputtering experiments, construct a digital model of microhabitat partitioning that includes shaded areas, dripping areas, and full-light areas. The environmental monitoring and quantification module is used to deploy a sensor network to collect and quantify environmental pressure characteristic data of each zone in real time. The environmental pressure characteristic data includes at least photosynthetically active radiation, soil moisture and drip scouring force. The species selection and optimization module has a built-in local native plant database, which is used to select suitable candidate species based on the environmental pressure characteristics data of each zone, and generate the optimal seed ratio for each zone by solving a multi-objective optimization problem. The precision seeding control module is used to receive the optimal seed ratio and the digital model of microhabitat zoning, generate operation files, and control the seeding equipment to carry out differentiated precision seeding in different zones based on real-time positioning information; The effect evaluation and feedback module is used to monitor the survival rate of vegetation after sowing and compare the measured data with the expected value. When the deviation exceeds the preset threshold, it triggers the correction of the input parameters of the species screening and optimization module.

[0016] In summary, this application includes at least one of the following beneficial technical effects: 1. By establishing a microhabitat zoning model based on the geometric parameters of photovoltaic modules and quantifying key environmental factors such as light, water, and erosion in each zone, a high degree of matching between native plant configuration and local ecological stress is achieved. Shade-tolerant species with high chlorophyll b content are used in the shade zone, fibrous root plants with strong soil-fixing ability are deployed in the drip zone, and drought-tolerant nitrogen-fixing legumes are introduced in the full-sun zone, which significantly improves the overall survival rate of planting and fundamentally solves the problem of "planting but not surviving, and surviving but not stable" caused by traditional uniform sowing throughout the field.

[0017] 2. This invention selects native grasses with erosion resistance and optimizes their sowing density to form a stable underground root network, significantly reducing the stripping and transport capacity of soil particles by surface runoff during the rainy season. Field measurements show a substantial reduction in soil erosion in drip irrigation areas, effectively avoiding the risks of hollowing and settlement of photovoltaic support foundations caused by soil erosion, and ensuring the long-term safe operation of the power station.

[0018] 3. This invention abandons the extensive restoration model that relies on human experience. Instead, it uses intelligent algorithms to generate precise, differentiated seed formulas, avoiding the over-sowing of ineffective species and frequent replanting later. Simultaneously, due to the high compatibility of vegetation with the environment, the need for artificial intervention such as irrigation and fertilization is reduced, significantly lowering the annual ecological restoration and maintenance costs for the entire site, thus achieving a synergistic improvement in both ecological and economic benefits. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the overall scheme for the intelligent configuration and restoration method of native species in photovoltaic power stations; Figure 2 This is a schematic diagram illustrating the principle of native plant adaptation screening based on microhabitat zoning and environmental stress feature vectors. Figure 3 This is a logical flowchart of the multi-objective collaborative configuration optimization phase; Figure 4This is a schematic diagram of the multi-level interaction and data flow of differentiated and precise seeding implementation and feedback adjustments in shaded areas, dripping areas, and full-sunlight areas. Detailed Implementation

[0020] Example 1

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0022] The intelligent configuration and restoration method for native species populations at photovoltaic power plants includes the following implementation steps: First, regarding step S1, namely the construction of the microhabitat partitioning model, its specific implementation process includes the following sub-steps: Step S101: Obtain spatial data and geometric parameters of the photovoltaic array. Accurate spatial layout information of the photovoltaic array is obtained by collecting as-built drawings of the site or using UAV oblique photogrammetry, and the ground clearance H and installation tilt angle θ of each row of photovoltaic modules are extracted from this data.

[0023] The height H above the ground refers to the vertical distance from the bottom edge of the module to the ground surface, and the value range is usually from 0.8 meters to 2.5 meters; the installation tilt angle θ is the angle between the plane of the module and the horizontal plane, and the value range is usually from 15° to 35°. This range covers the engineering practice range of typical fixed photovoltaic brackets in Northwest, North and East my country.

[0024] Step S102: Determine the geographical latitude of the site. Measure the geographical coordinates of the site's center point using a GPS positioning device, or obtain the local geographical latitude δ from the site design data. This latitude value will be used in subsequent shadow length calculations.

[0025] Step S103: Calculate the shadow boundary under the most unfavorable lighting conditions throughout the year. Using the lowest solar altitude angle at noon on the winter solstice as the benchmark, the solar declination angle φ is taken as -23.44°. The formula for calculating the solar altitude angle h is h = 90° - |δ - φ|. Since δ is greater than |φ| in the Northern Hemisphere, it can be simplified to h = 90° - δ + φ, where φ is a negative value.

[0026] The projection length L of the photovoltaic module on flat ground is determined by the following formula: L = H / tan(h) = H / tan(90° - δ + φ). This length L represents the distance extending northward from the vertical projection point along the lower edge of the module, and this projection area is defined as the shadow area. This means that at noon on the winter solstice, this area is completely shaded by the module and can be used as a representative area of ​​the most severe light conditions throughout the year for subsequent assessment of light stress.

[0027] Step S104: Delineate the dripping zone. The delineation of the dripping zone is based on hydrodynamic splashing experiments. Under laboratory conditions, the maximum rainfall intensity is simulated, for example, set at 50 mm / h. Using high-speed imaging technology combined with particle image velocimetry, the maximum splash radius of rainwater dripping from the lower edge of the component onto surface soil particles is determined.

[0028] Multiple field measurements show that this radius value is mainly concentrated between 15 cm and 25 cm. Therefore, the annular strip area formed by extending 20 cm outward from the vertical projection line at the bottom edge of the component is defined as the dripping area.

[0029] It should be noted that the 20 cm value is an empirical value based on typical conditions. In practical applications, it can be adjusted using the same experimental method according to the local extreme rainfall intensity and soil texture.

[0030] Step S105: Delineate the all-light zone. Within the site area, after deducting the already designated shaded and dripping areas, all remaining surface areas are uniformly classified as all-light zones.

[0031] Based on the physical structure of photovoltaic modules and the environmental effects they cause, the entire site surface is divided into three types of microhabitat units: shaded area, dripping area, and full-sunlight area.

[0032] Step S106: Zoning and Vectorization. Using the Geographic Information System platform, the spatial boundaries of the defined shaded areas, dripping areas, and full-light areas are vectorized to generate an electronic map containing multiple polygon layers.

[0033] Each polygon is assigned a unique identifier, and its attribute table records its partition type, precise area, center geographic coordinates, and topological relationships with adjacent polygons. This digital partition map serves as the spatial data foundation for the entire method, used for subsequent environmental sensor placement planning and precision seeding path navigation, ensuring all operations are accurate and precise.

[0034] Thus, the entire process of microhabitat zoning starts from the geometric parameters of photovoltaic modules, and through scientific calculations and experimental data, transforms abstract differences in light intensity and rain impact into concrete and measurable spatial polygons, providing a high-precision operational blueprint for subsequent ecological restoration.

[0035] The specific implementation process for step S2, namely the quantization and construction of the environmental pressure feature vector, includes the following sub-steps: Step S201: Deploy a photosynthetically active radiation monitoring network. In both the shaded and fully lit areas, deploy a quantum sensor array with a grid density of 50 meters by 50 meters. Select a LI-190R sensor with a spectral response range of 400 nm to 700 nm and a measurement accuracy of ±5%.

[0036] Each sensor is connected to the data acquisition unit via a signal line. The data acquisition unit is equipped with a real-time clock and a storage module, and the sampling frequency is set to record and store data once every 10 minutes.

[0037] The monitoring period should fully cover the main growing season of local plants, usually from April to October, and ensure that the cumulative number of valid data days is no less than 180 days.

[0038] All raw photosynthetically active radiation data are collected by the data acquisition device and then transmitted to the on-site industrial control computer or cloud server via wired or wireless means for subsequent analysis and processing.

[0039] Step S202: Process photosynthetically active radiation (PANR) data. The collected raw PANR data undergoes time integration processing. Specifically, instantaneous values ​​taken every 10 minutes are accumulated on the time axis to obtain the daily cumulative PANR value for each monitoring point. These values ​​are then further accumulated to obtain the cumulative PANR value for the entire growing season. At the same time, the daily data are arithmetically averaged to obtain the daily average photosynthetically active radiation value.

[0040] Subsequently, spatial interpolation methods such as Kriging interpolation or inverse distance weighted interpolation are used to obtain the data for all monitoring points. Based on the values, a spatial distribution map of cumulative photosynthetically active radiation in the shaded and fully lit regions is generated.

[0041] Statistical results show that the seasonal cumulative photosynthetically active radiation value in the shaded area is usually 15% to 25% of that in the full-light area. This ratio will serve as a key input threshold for screening shade-tolerant species in subsequent steps.

[0042] Step S203: Set up soil moisture monitoring points and collect data. In the shaded area, dripping area and full-sunlight area, set up time domain reflectometer probes according to the principle of representativeness. The probes are vertically inserted into the soil layer to a depth of 0 to 30 cm, which covers the main distribution area of ​​most herbaceous plant roots.

[0043] Connect the probe to the data acquisition device and set the recording interval to automatically record soil volumetric water content every 2 hours. The data is stored in real time. Simultaneously, undisturbed soil samples are collected near each monitoring point using a soil auger, sealed, and brought back to the laboratory for particle size analysis to determine the mass percentages of sand, silt, and clay particles in the soil. This basic data will be used for subsequent calculations of the soil moisture availability index.

[0044] Step S204: Calculate the soil moisture availability index. Based on the soil particle size analysis results, using a soil transfer function model such as the ROSETTA model, input the percentages of sand, silt, and clay particles to calculate the field capacity at each monitoring point. and moisture content at wilting point Field holding capacity refers to the maximum water content that soil can retain after adequate drainage, while wilting point water content refers to the critical water content at which plants can no longer absorb water from the soil and will permanently wilt. Subsequently, the soil moisture availability index (SMEI) for each sampling time was calculated using the following standardized formula: in This represents the volumetric water content measured in real time. The index value ranges from 0 to 1, with lower values ​​indicating more severe water stress on the plant. For each monitoring point, the SMEI value at all times throughout the monitoring period is arithmetically averaged to obtain the representative SMEI value for that point.

[0045] According to measured data, in shaded areas, rainwater is difficult to reach due to component shading, resulting in an average SMEI generally below 0.3; in fully lit areas, the average SMEI can drop below 0.2 under no irrigation conditions. These typical values ​​will be used for subsequent screening of drought-resistant species.

[0046] Step S205: Deploy the scouring force monitoring device in the dripping area. Within the dripping area, select representative locations to install embedded miniature flow meters.

[0047] This flow meter integrates a piezoelectric thin-film sensor and a data logging module. The sensor's sensing surface is buried 5 millimeters below the surface to ensure direct sensing of the velocity of water flow in a thin layer at the ground. The device is connected to a data acquisition unit via cable and a trigger sampling mode is set.

[0048] At the same time, tipping bucket rain gauges are deployed within the station, and the output signals of the rain gauges are also connected to the same data acquisition unit.

[0049] Step S206: Real-time monitoring of water flow velocity during rainfall events. The data acquisition unit continuously monitors the rain gauge signal. When a typical rainfall event with a rainfall intensity of 20 mm / hour or higher is detected, the flow meter is automatically triggered to sample the water flow velocity u at a high frequency of 100 Hz.

[0050] During each rainfall event, sampling continues until the rainfall intensity falls below a threshold. The sampled data is stored in real-time in the data recording module. For each sampling moment, the scouring force intensity F is expressed as the momentum flux per unit area, calculated using the following formula: Where ρ is the density of water, taken as 1000 kg per cubic meter. The instantaneous scouring force at each sampling time is calculated using this formula.

[0051] For each rainfall event, its peak scour force or average scour force can be calculated. Over the entire monitoring period, the average of the peak scour forces from all rainfall events is taken as the representative scour force at that point in the dripping area. Actual measurement data shows that under a rainfall intensity of 50 mm / hour, the peak scour force at the center of the dripping area can reach 1.2 to 2.5 Newtons per square meter, while the scour force in the full-sunlight area is usually less than 0.3 Newtons per square meter during the same period. This value will serve as a direct basis for subsequent screening of scour-resistant species.

[0052] Step S207: Integrate and form the environmental pressure characteristic vectors for each zone. For shaded and fully lit areas, calculate the seasonal cumulative photosynthetically active radiation values ​​of all monitoring points. Calculate the arithmetic mean to obtain the representative value of the partition. The representative SMEI values ​​of all monitoring points are arithmetically averaged to obtain the representative SMEI of the zone; the scouring force F of the two zones is uniformly set to 0.

[0053] For the dripping area, the representative scouring force at all flowmeter monitoring points within the area will be used. The arithmetic mean was calculated to obtain the representative F for this zone; simultaneously, the average SMEI values ​​of the soil moisture monitoring points deployed within the dripping zone were used as the SMEI for this zone, and the average values ​​of the photosynthetically active radiation monitoring points deployed within the dripping zone were also used as the SMEI. The average is used as the partition. Finally, a three-dimensional environmental pressure feature vector is constructed for each microhabitat partition. .

[0054] This feature vector fully quantifies the main environmental stresses experienced by each partition and will be directly used to drive the native plant adaptation screening logic in subsequent steps.

[0055] In summary, step S2, through the aforementioned monitoring and calculations, transforms the complex micro-habitat environmental conditions within the photovoltaic power station into a quantified feature vector, providing a scientific basis and data support for the subsequent precise matching of species and environment.

[0056] For step S3, namely the screening of native plant suitability, the specific implementation process includes the following sub-steps: Step S301: Construct a local native plant resource database. First, collect a list of herbaceous and shrub species that have been successfully used in ecological restoration projects in the target area over the past 10 years.

[0057] For each species, key trait data were determined or obtained using the following methods: morphological parameters, including plant height and crown width at maturity, were obtained through field measurements or by consulting relevant literature; physiological and ecological indicators, including light compensation point and water use efficiency, were measured using a portable photosynthesis meter under controlled light and water conditions.

[0058] Reproductive characteristics include seed weight per thousand seeds and germination rate. Seed weight per thousand seeds was obtained by weighing 1000 pure seeds using a laboratory electronic balance. Germination rate was obtained by statistical analysis of germination tests conducted in a constant temperature incubator. Root system architecture data includes fibrous root density, taproot depth, and root tensile strength. Fiber root density was obtained by excavation to obtain the total length of fibrous roots per unit volume of soil and analyzed using a root scanner. Taproot depth was measured directly after excavation. Root tensile strength was tested on individual roots using a universal testing machine until fracture, and the maximum tensile force was recorded.

[0059] All data is stored according to a relational database structure. Each species' fields include Latin name, Chinese name, family and genus, functional group classification, and quantitative values ​​of the above traits, thus completing the local native plant resource database.

[0060] Step S302: Set the shade tolerance screening threshold for the shaded area and perform the screening. Based on the environmental pressure characteristics of the shaded area obtained in step S2, the photosynthetically active radiation in this area is only 15% to 25% of that in full-light areas. Therefore, the shade tolerance screening conditions are set as follows: Under shading conditions where the photosynthetically active radiation is 20% of that in full-light areas, the net photosynthetic rate maintenance rate of candidate species shall not be less than 60%, that is, the ratio of the net photosynthetic rate measured under shading to the net photosynthetic rate measured under full light shall not be less than 0.6.

[0061] Simultaneously, the ratio of chlorophyll b to chlorophyll a must be greater than 1.2. This ratio is calculated after measuring the chlorophyll content of leaves using spectrophotometry. In the resource database, species that simultaneously meet both of these conditions are selected using SQL queries. Taking *Cephalotaxus fortunei* and *Adiantum capillus-veneris* as examples, their measured chlorophyll b to chlorophyll a ratios are 1.35 and 1.42 respectively, both meeting the selection requirements.

[0062] Step S303: Set the erosion resistance screening threshold for the dripping area and perform the screening. Based on the erosion force intensity of the dripping area obtained in step S2, its peak erosion force can reach 1.2 to 2.5 Newtons per square meter. Therefore, a dual threshold for erosion resistance is set: root density not less than 8 km / m³ and root tensile strength not less than 0.8 Newtons. Root density is determined by the method described in step S301, and root tensile strength is also obtained through a tensile test.

[0063] In the resource database, species that simultaneously meet both conditions are selected using SQL queries. Taking *Achnatherum splendens* as an example, its measured root density is 9.2 km / m³, and its root tensile strength is 1.1 Newtons, which fully meets the requirements.

[0064] Step S304: Set the drought tolerance screening threshold for full-sunlight areas and perform the screening. Based on the soil moisture availability index obtained in step S2, the index can drop below 0.2 under no irrigation conditions. Therefore, the drought tolerance judgment criteria are set as follows: when the soil moisture availability index is not greater than 0.25, the plant can still maintain a stomatal conductance greater than 0.1 mol / m² / s and a relative leaf water content greater than 70%.

[0065] Stomatal conductance was measured under controlled moisture conditions using a portable photosynthesis meter or stoma meter. The relative water content of the leaves was determined by the weighing method, which involved collecting leaves, weighing their fresh weight, soaking them in water until saturated, weighing them again, drying them, weighing them again, and then calculating the relative water content.

[0066] In the resource database, species that meet this physiological requirement are selected using SQL queries. Taking *Amorpha fruticosa* and *Amorpha tinctoria* as examples, both can maintain stable physiological activity under this stress condition. Both are leguminous plants with nitrogen-fixing ability in root nodules, which can increase the total nitrogen content of the soil by 0.1 to 0.3 grams per kilogram of soil per year.

[0067] Step S305: Integrate and output the candidate species list and trait compliance matrix. Summarize all species selected from the above three partitions and their trait compliance markers to generate a three-dimensional matrix.

[0068] The rows of the matrix represent each candidate species, and the columns represent three partitions. Each cell records the achievement status of each key trait of the species in that partition, for example, indicated by achievement or non-achievement.

[0069] This matrix will serve as the direct input to the multi-objective optimization algorithm in step S4, used to determine the optimal seed ratio for each partition.

[0070] In summary, by establishing a local native plant resource bank and setting quantitative screening thresholds based on the environmental pressure characteristics of each zone, the steps accurately screen candidate species suitable for different microhabitats from a large number of native species, providing a reliable species bank and trait data for subsequent multi-objective optimization.

[0071] For step S4, multi-objective collaborative configuration optimization, its specific implementation process includes the following sub-steps: Step S401: Define decision variables and basic parameters. Assume there are n candidate species selected in step S3, and the entire site is divided into three microhabitat zones, represented by k=1, 2, and 3 respectively, representing the shaded area, dripping area, and full-light area. Calculate the proportion of each zone's area to the total area of ​​the entire site from the digital zoning map generated in step S1, denoted as A1, A2, and A3, satisfying A1+A2+A3=1.

[0072] Define the decision variable as the seeding rate per unit area for each species within each partition, using... The seeding amount of species j in partition k is expressed in grams per square meter. All decision variables form an n x 3 matrix. These variables will be the direct controls for subsequent optimization.

[0073] Step S402: Set optimization constraints. The optimization problem must satisfy the following three types of constraints: Total seeding limit: The total seeding amount used in the entire field must not exceed the preset limit. Typically, 20 grams per square meter is taken, that is... Minimum Coverage Constraint for a Single Species: To ensure the uniformity of species composition within each partition, if the total seeding density of a partition k is... If the value is greater than zero, then the sowing ratio of any species j within that partition is... Not lower than the minimum ratio , Let's take 0.05. That is, for all k and j, if... Then there must be .

[0074] Species compatibility constraints: Construct a species compatibility matrix C, whose elements The determination is based on whether there is allelopathic inhibition or competitive exclusion between species j and l. If inhibition exists, then... ,otherwise This matrix can be obtained by consulting relevant literature or by conducting prior potted plant confrontation experiments. Simultaneous occurrences are not allowed within the same partition. For each partition k, if the two species are... and Then it must satisfy .

[0075] Step S403, construct the first objective function: maximize the weighted sum of the expected survival rates of vegetation in each zone. First objective function The calculation formula is: in The weighted average survival rate of partition k is determined by the sowing weight of each species within that partition. With predicted survival rate Weighted average yields: In the formula The predicted survival rate of species j under environmental pressure in partition k is calculated. This predicted value is obtained through the following method: based on the existing species trait data in the database of step S3 and the measured survival rate records of species under similar environments in historical ecological restoration projects, a multiple regression model is established. Specifically, the light compensation point, water use efficiency, fibrous root density, and root tensile strength of the species are selected as independent variables, while the photosynthetically active radiation in the environmental pressure feature vector of partition k is also included. Soil moisture availability index (SMEI) and scour force (F) were used as covariates, with the measured survival rate as the dependent variable. Multiple linear regression or random forest regression methods were employed to fit the regression equation. After model training, for any species j and region k, the corresponding trait values ​​and environmental factors could be substituted into the equation to calculate... If sufficient historical data is lacking, the survival rate of similar species in similar habitats in the literature can be used as a reference value.

[0076] Step S404: Construct the second objective function: minimize the total soil erosion at the entire site. The modified general soil loss equation RUSLE, coupled with the root system soil-fixing factor, is used. The calculation formula is as follows: The meanings and determination methods of each factor in the formula are as follows: R is the rainfall erosivity factor, measured in megajoules per millimeter per hectare per hour per year. It can be determined using an empirical formula based on multi-year average annual rainfall or rainfall event data from meteorological stations in the area where the station is located. Calculate, where E is the total kinetic energy of a single rainfall event. The maximum 30-minute rainfall intensity for this event can also be determined by referring to the R value given in the local soil erosion survey report or by using the contour map in the "Soil Erosion Classification and Grading Standard".

[0077] K is the soil erodibility factor, expressed in tons-hectare-hours per hectare-megajoule-mm. Based on the percentage of soil sand, silt, and clay particles and the organic matter content determined in step S2, K is expressed using a Wischmeier nomograph or formula. The calculation is performed, where M is the product of (silt + very fine sand) percentage and (100 - clay percentage), OM is the organic matter percentage, s is the soil structure grade, and p is the profile permeability grade.

[0078] LS is the slope length and slope factor, which is dimensionless. The slope length λ and slope θ are extracted from the digital elevation model (DEM) of the site, according to the formula... Calculate, where m is the slope length index, which is generally taken as 0.5.

[0079] is the vegetation cover management factor for zone k, dimensionless. It is closely related to the root density of species within the zone and is calculated as a weighted average of the fibrous root densities of all species in that zone. ( (where the root density of species j is given by an empirical formula) The calculation shows that α is an empirical coefficient, which can be taken as 0.05.

[0080] P is a dimensionless factor representing soil and water conservation measures. If no special soil and water conservation measures are taken, P is taken as 1; if measures such as contour farming and terraced fields are taken, the value is taken according to the corresponding specifications.

[0081] Step S405: Construct the third objective function: maximize the plant community functional diversity index FD. Select functional traits that reflect the complementarity of resource utilization, such as specific leaf area, root depth, and nitrogen fixation capacity.

[0082] These trait values ​​for each species are extracted from the database in step S3, forming an n x m trait matrix (where m is the number of traits). First, the trait values ​​are standardized to eliminate the influence of dimensions; range standardization can be used: for the t-th trait, its standardized value is... .

[0083] Then, taking each species as a point, calculate the volume of the convex polyhedron formed by all species in the m-dimensional trait space. This volume is the FD exponent. The calculation can be performed using the Quickhull algorithm or by calling the dbFD function in the R language's FD package.

[0084] The higher the FD value, the higher the community functional diversity and the stronger the complementarity of resource utilization.

[0085] Step S406: The non-dominated sorting genetic algorithm NSGA-II is used to solve the multi-objective optimization problem. Algorithm parameters are set as follows: population size N = 200, maximum number of generations G = 500, crossover probability Pc = 0.9, mutation probability Pm = 0.1. Encoding method: The decision variables are... Arrange them in order as a real vector to form a chromosome.

[0086] During population initialization, N feasible solutions satisfying the constraints in step S402 are randomly generated. For each individual in each generation of the population, the constraints are first checked. If the constraints are not met, the fitness of the individual is reduced using a penalty function method or it is directly eliminated during selection.

[0087] For feasible individuals, calculate three objective function values ​​f1, f2, and f3 respectively. Then, perform fast non-dominated sorting to determine the non-dominated level of each individual based on the objective function values; for individuals within the same non-dominated level, calculate their crowding distance.

[0088] Based on the principles of prioritizing non-dominated levels and prioritizing those with greater crowding distance within the same level, a tournament selection method is used to select parent individuals. Offspring individuals are generated through simulated binary crossover and polynomial mutation. The parent and offspring populations are merged, and non-dominated sorting and crowding calculations are performed again. The top N individuals are selected for the next generation. This process is iterated until the maximum number of generations G is reached, ultimately yielding several non-dominated solutions on the Pareto optimal frontier.

[0089] Step S407: Select the final seed formula from the Pareto optimal solution set. For each non-dominated solution on the Pareto front, normalize its three objective function values. The normalization uses the mini-maximum method: for each objective f, take the minimum value of that objective in the current solution set. and maximum value Then the normalized value Then, the weighted summation method is used to calculate the comprehensive evaluation value. Where f'2 is the amount of soil erosion, so we take 1-f'2 to make its direction consistent, and the larger the better.

[0090] The weighting coefficients can be set according to engineering requirements. For example, the survival rate weight ω1=0.4, the soil and water conservation weight ω2=0.3, and the functional diversity weight ω3=0.3. The non-inferior solution with the highest comprehensive evaluation value is selected as the final solution.

[0091] This solution gives all The value represents the absolute seeding quantity of each species within each partition. This allows for the calculation of the total seeding quantity for each partition. The seed weight ratio for each zone, W1:W2:W3=T1:T2:T3, can be further refined to a detailed list of the specific sowing quantities for each species in each zone.

[0092] Step S408: Output the optimization results. Store the final seed formula as a data file, including at least the zone identifier, species name, and seeding rate (grams per square meter). This file is transmitted to the seeding operation control system in step S5 to guide drones or mechanical spraying equipment in performing differentiated precision seeding. Simultaneously, retain the parameter settings and intermediate results from the optimization process for subsequent effect evaluation and feedback adjustments.

[0093] In summary, step S4, by constructing a multi-objective optimization model and solving it using a non-dominated sorting genetic algorithm, obtained a differentiated seed formula that precisely matches the environmental pressure of each microhabitat, providing a scientific basis for maximizing vegetation survival rate, soil and water conservation effect, and ecological function diversity.

[0094] For step S5, namely differentiated precision seeding, its specific implementation process includes the following sub-steps: Step S501: Obtain the seed formula and zoning map. Obtain the final seed formula from the optimization results output in Step S4. This formula should include the specific sowing amount for each species within each microhabitat zone, in grams per square meter. Simultaneously, extract the spatial boundary vector layers and their attribute data for shaded areas, drip zones, and full-sunlight zones from the digital zoning map generated in Step S1. Associate the seed formula data with the zoning map to generate a job file containing zone identifiers, spatial extents, and corresponding sowing amount lists. This file will be used to guide the control of subsequent sowing equipment.

[0095] Step S502: Prepare the mechanical spraying equipment. A tractor-tethered, zone-specific mechanical spraying system is used. This equipment must be equipped with a high-precision GPS positioning module, employing real-time dynamic differential (RTK) technology, achieving a horizontal positioning accuracy of ±2 cm to ensure accurate identification of the current zone. The equipment is equipped with multiple independent seed storage compartments, each storing the seed mixture for one zone. A variable valve assembly controlled by a solenoid valve is installed at the outlet of each storage compartment, allowing for real-time adjustment of the seed flow rate based on control signals. The equipment also includes a pressure pump and spraying nozzles. The spraying pressure should be maintained between 0.3 MPa and 0.5 MPa to ensure the mixture of seeds, water-retaining agent, and adhesive is evenly sprayed and adheres to the soil surface.

[0096] Step S503, Preparations before mechanical hydroseeding. Import the operation file and microhabitat zoning vector map generated in step S501 into the vehicle-mounted control system. The control system has a built-in geographic information system engine that can receive GPS positioning signals in real time and determine which zoning zone the current location of the equipment belongs to based on the zoning map. Simultaneously, the control system sets the zoning type corresponding to each seed storage bin and the target seeding rate for each species within that zoning zone. The system automatically calculates the required seed flow control parameters based on the target seeding rate and the preset operating width and travel speed.

[0097] Step S504: Perform mechanical spraying. The tractor travels along a pre-planned path, and the onboard control system continuously reads GPS location information and determines the zone it belongs to in real time. When the tractor enters a zone, the system automatically activates the solenoid valve of the corresponding seed storage compartment and dynamically adjusts the valve opening according to the current travel speed and target seeding amount to ensure that the actual seeding amount per unit area meets the formula requirements. During spraying, the pressure pump maintains the set pressure to ensure that the seed mixture is sprayed evenly. The system automatically records the actual sowing area of ​​each zone, the seed consumption of each sub-compartment, and equipment operating parameters, generating a mechanical spraying operation log.

[0098] Step S505: Prepare the drone precision seeding system. A six-rotor drone platform is used, equipped with a four-channel independent seeding device. Each channel corresponds to a seed formula component; for example, seed mixtures for shaded areas, drip irrigation areas, and full-sunlight areas can be stored in three separate channels, with the remaining channel reserved or used for spreading auxiliary materials such as water-retaining agents. The seeding device incorporates a rotating disk driven by a stepper motor. Pulse Width Modulation (PWM) signals precisely control the motor speed, thereby adjusting the seed output per second, achieving a seeding accuracy error of less than 10 centimeters. The drone also integrates a wind speed sensor for real-time wind speed and direction measurement, providing data for flight control and landing point compensation.

[0099] Step S506: Drone Seeding Route Planning. Using flight path planning software, import the site's Digital Elevation Model (DEM) and microhabitat zoning vector map. Based on the drone's flight performance parameters, the software generates a reciprocating route covering the entire field. The route design must ensure that the drone can automatically switch seeding channels according to zoning boundaries during flight. Set the flight altitude to 5 meters and the flight speed to 3 meters per second. The route should cover all zoning zones and ensure appropriate overlap between adjacent routes to prevent missed seeding.

[0100] Step S507: Perform drone seeding operation. The drone flies autonomously along the planned route, and the onboard flight control system acquires its GPS position in real time and matches it with the zone map. When the drone enters a zone, the flight control system sends a PWM control signal to the seeding device in the corresponding channel according to the current zone type, starts the stepper motor and adjusts its speed according to the preset seeding amount, so that the seeds are discharged at a constant rate. The wind speed sensor collects data in real time, and the flight control system automatically adjusts the flight trajectory or compensates for the seeding landing point according to the crosswind strength to ensure that the seeds fall accurately into the target zone. During the operation, the drone records the flight trajectory, the seed consumption of each channel, and the operation parameters, generating a drone seeding operation log.

[0101] Step S508: Post-operation data summary. Regardless of whether mechanical hydroseeding or drone seeding is used, the generated operation log should be uploaded to the data center after the operation is completed. The log should include information such as the actual seeding area of ​​each zone, the seed consumption of each species, the operation time, and the equipment operating status. This data will be used to evaluate the subsequent vegetation restoration effect and, if necessary, as a basis for adjusting seeding parameters and optimizing the formula.

[0102] Thus, step S5 achieves differentiated and precise sowing based on microhabitat zoning through the two technical approaches described above, ensuring that each zoning receives a seed formula that matches its environmental stress, laying a solid foundation for subsequent vegetation restoration and ecological repair.

[0103] In addition, step S6, which addresses the post-monitoring and feedback adjustment mechanism, includes the following sub-steps in its implementation process: Step S601: Set monitoring time points and evaluation indicators. After sowing, conduct three phased monitoring sessions on day 30, day 60, and day 90. Each monitoring session should evaluate the vegetation cover, average plant height, and survival density of each microhabitat zone.

[0104] Vegetation cover refers to the percentage of the vertical projection area of ​​vegetation to the area of ​​the sample plot. Average plant height is obtained by randomly measuring at least 10 plants. Survival density refers to the number of surviving plants per unit area, expressed as plants per square meter. The actual survival rate obtained from monitoring is compared with the weighted average survival rate predicted for that area in step S4. Compare them.

[0105] Step S602, ground quadrat survey. Within each zone, set up no fewer than three 1-meter by 1-meter quadrats using a random or systematic sampling method.

[0106] The quadrat location must be representative, avoiding areas with obvious anomalies. Investigators enter the quadrat, manually record the species and quantity of all surviving plants within the quadrat, and calculate the survival density and vegetation cover of the quadrat.

[0107] Coverage can be estimated using visual estimation or grid method. The average value of multiple quadrats is used as the measured coverage, plant height and survival density of the area, and the measured survival rate, i.e. the ratio of survival density to sowing density, is calculated based on this.

[0108] Step S603: UAV multispectral remote sensing monitoring. Simultaneously or at approximately the same time as each ground survey, a UAV equipped with a RedEdge-MX multispectral camera is used for low-altitude remote sensing flight. The flight altitude is set between 50 and 100 meters, with both forward and lateral overlap rates not less than 70%, acquiring multispectral images covering the entire field, including blue, green, red, red-edge, and near-infrared bands.

[0109] After radiometric calibration and geometric correction of the imagery, the Normalized Difference Vegetation Index (NDVI) is calculated using the following formula: Where NIR is the reflectance in the near-infrared band and Red is the reflectance in the red band.

[0110] Step S604, Vegetation cover inversion based on machine learning. Using ground quadrat survey data as training samples, the measured cover within the quadrat is used as the output, and the NDVI value and other spectral indices such as ratio vegetation index (RVI) and enhanced vegetation index (EVI) at the corresponding location of the quadrat are used as input features to construct a regression model.

[0111] The model can employ random forest regression or support vector regression, with optimal parameters selected through cross-validation. After model training, the NDVI and other features of all pixels in the entire field are calculated and substituted into the model to obtain the predicted vegetation cover value for each pixel, thereby generating a vegetation cover distribution map for the entire field. The average cover of each zone is extracted from the distribution map and verified and fused with the ground quadrat results to obtain the final measured cover and survival rate of each zone.

[0112] Step S605, Deviation Judgment and Trigger Adjustment. The measured survival rate of each partition obtained in step S604 is compared with the predicted survival rate of the corresponding partition in step S4. The relative deviation value δ is calculated by comparing the measured value with the predicted value and then dividing the predicted value by the predicted value. If δ exceeds a preset deviation threshold, such as 15%, it is considered that there is a significant difference between the current model prediction and the actual situation, and the feedback adjustment mechanism needs to be activated.

[0113] Step S606: Retrospectively analyze the sources of error. First, check the environmental parameter acquisition system, including whether the photosynthetically active radiation sensor is drifting or contaminated, whether the time domain reflectometer probe is in good contact with the soil, and whether the rain gauge and current meter are working properly.

[0114] Troubleshooting can be done by checking for outliers in the data logs, comparing data with backup sensor data, and conducting on-site inspections.

[0115] Next, check the species trait database to confirm whether the species trait data used in step S3 is outdated or has been entered incorrectly. This can be verified by consulting the latest literature or re-measuring some traits. Also, check whether the regression model used to predict survival rate in step S403 needs recalibration.

[0116] Step S607: Correct parameters and re-optimize. Based on the retrospective analysis results, correct problematic environmental parameter thresholds or sensor data, such as removing abnormal monitoring data, recalibrating sensors, and updating the corresponding values ​​in the species trait database.

[0117] Then, using the corrected data as input, steps S2 to S4 are re-executed, namely, requantifying the environmental stress feature vector, re-screening species, and re-running the multi-objective optimization algorithm to generate an updated seed formula.

[0118] Step S608: Generate an iterative optimization report. Record and archive the entire process of this feedback adjustment, including deviation analysis results, error source investigation, parameter corrections, and the newly generated seed formula. This report can be used to guide subsequent sowing operations and provide experience for ecological restoration at similar sites in the future.

[0119] Thus, through the aforementioned post-monitoring and feedback adjustment mechanism, dynamic evaluation of the restoration effect and continuous optimization of model parameters are achieved, ensuring that the seed formula can be continuously iterated and improved with environmental changes and data accumulation, ultimately achieving the best ecological restoration effect. Example 2

[0120] In another application scenario, for photovoltaic power stations in hilly areas of southern China with high rainfall intensity, such as areas with an average annual rainfall of 1200 mm, the present invention makes adaptive adjustments to the relevant steps, and the specific implementation process is as follows: Adjust the width of the drip zone. Based on the local 50-year return period rainfall intensity, i.e., 100 mm / hour of rainfall, repeat the hydrodynamic splash experiment in step S104 to determine the maximum splash radius of rainwater dripping from the lower edge of the component onto surface soil particles. The measured results show that under this extreme rainfall intensity, the splash radius extends to 30 cm. Therefore, the annular strip extending 30 cm outward from the vertical projection line of the lower edge of the component is designated as the drip zone to accommodate stronger rainwater erosion.

[0121] Upgrade the scour force monitoring equipment in the dripping area. Replace the embedded micro flow meter in step S205 with a high-frequency pressure sensor array. This array consists of multiple piezoelectric or piezoresistive pressure sensors, with the sensing surface also buried 5 mm below the surface, enabling rapid response to changes in water pressure. The sensor array is connected to a data acquisition unit via cable, increasing the sampling frequency to 1000 Hz (1 kHz) to capture the instantaneous peak scour force during short-duration heavy rainfall. The data acquisition unit simultaneously records the signal from the tipping bucket rain gauge, automatically triggering sampling when the rainfall intensity reaches 20 mm per hour. For each sampling time, the scour force intensity F is still calculated as F = ρ·u 2 The calculations were performed, where the water flow velocity *u* was derived from the dynamic pressure measured by a pressure sensor, and the water density *ρ* was taken as 1000 kg / m³. Actual measurement data showed that under a rainfall intensity of 100 mm / h, the peak scouring force at the center of the dripping area could reach 3.5 N / m².

[0122] Raise the erosion resistance screening threshold. Based on the higher erosion force value obtained in step 2, adjust the erosion resistance index of the drip zone to: root density not less than 10 km / m³, and root tensile strength not less than 1.2 N. Based on the native plant resource database established in step S301, use SQL queries to screen species that meet the new threshold. Taking Bermuda grass as an example, its rhizomes are well-developed, the measured root density is 11.3 km / m³, and the root tensile strength is 1.35 N, which fully meets the requirements. Therefore, it is selected as a candidate species for the drip zone.

[0123] The rainfall erosivity factor in the soil erosion equation was recalibrated. For this hilly southern region, rainfall intensity formulas and rainfall kinetic energy data were obtained from local meteorological stations to calculate the total kinetic energy E and the maximum 30-minute rainfall intensity for each rainfall event. The product of the two factors is summed to obtain the new rainfall erosivity factor R. The recalibrated R value is then substituted into the modified general soil loss equation RUSLE in step S404 for the calculation of the soil and water loss objective function f2.

[0124] Finally, the adjusted parameters were input into the multi-objective optimization model in step S4 to re-solve the seed formula. The optimization results showed that the sowing weight W_drop in the drip zone was increased to 0.4, ensuring that the root network integrity could still be maintained under extreme rainfall conditions. Example 3

[0125] In application scenarios in arid desert areas, such as the Turpan region of Xinjiang, where the average annual rainfall is less than 50 mm, this invention adaptively adjusts relevant steps to enhance the synergy between drought resistance screening and water management in all-sunlight areas. The specific implementation process is as follows: Adjust soil moisture monitoring parameters. Extend the monitoring period for the soil moisture availability index from the growing season to the whole year, as the winter freeze-thaw process also affects seed germination and seedling survival in the following year. Simultaneously, increase the insertion depth of the time domain reflectometer probe from 30 cm to 50 cm to cover the main water absorption layers of deep-rooted plants, ensuring accurate monitoring of deep soil moisture conditions. Maintain a data acquisition frequency of once every 2 hours, and simultaneously record soil temperature for freeze-thaw process analysis.

[0126] Expanding the candidate species list for full-sunlight zones and verifying drought tolerance: Camel thorn was added to the full-sunlight zone candidate species list. Camel thorn has a root system reaching depths of up to 3 meters, enabling it to utilize deep soil moisture. Through controlled water experiments, it was determined that when the soil moisture availability index dropped to 0.15, camel thorn still maintained normal physiological activities, including a stomatal conductance greater than 0.1 mol / m² / s and a relative leaf water content greater than 70%. Therefore, it was included in the full-sunlight zone candidate species list.

[0127] Water conservation constraints are added and optimization weights are adjusted. In the optimization constraints set in step S402, a fourth constraint is introduced: the maximum total seeding amount X_max is reduced from 20 grams per square meter to 15 grams per square meter to conserve precious water resources and reduce competition between seeds and water. Simultaneously, in step S403, the first objective function f1 = ΣA... k ·S k In this context, since the area of ​​a photovoltaic power station in a desert region often exceeds 70% due to its full-sunlight coverage, its weight A is... kThe primary optimization direction is to prioritize ensuring the expected survival rate in the fully lit region. The remaining objective functions and solution algorithms remain unchanged, and the optimization is rerun to obtain the seed formula.

[0128] Improved sowing methods were employed to enhance water use efficiency. Mechanical hydroseeding was used, with a water-retaining agent added to the seed mixture. Polyacrylamide was chosen as the water-retaining agent, applied at a rate of 2 grams per square meter, mixed with seeds, adhesive, and water before being sprayed evenly. The water-retaining agent can absorb and retain hundreds of times its own weight in water, forming a miniature water reservoir around the seeds, significantly improving water use efficiency during seed germination. The spraying pressure was maintained at 0.3 to 0.5 MPa to ensure the mixture adhered evenly to the soil surface.

[0129] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

[0130] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention. Therefore, the embodiments should be regarded as exemplary and non-limiting in all respects.

[0131] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A method for intelligent configuration and restoration of native species populations in photovoltaic power plants, characterized in that, Includes the following steps: Obtain the geometric parameters of the photovoltaic array, which include at least the height above the ground and the installation tilt angle; Based on the ground clearance and installation tilt angle, combined with the solar altitude angle, the boundary of the shadow area formed by the photovoltaic module is calculated; The maximum splash radius of raindrops under rainfall conditions was determined based on hydrodynamic sputtering experiments, and the boundary of the dripping area located at the lower edge of the component was delineated based on the maximum splash radius. The area within the station excluding the shaded area and the dripping area is divided into a fully lit area, and a microhabitat partitioning model including the shaded area, the dripping area and the fully lit area is constructed. Environmental pressure characteristic data of the shaded area, dripping area and full-light area were collected and quantified respectively. The environmental pressure characteristic data included at least photosynthetically active radiation, soil moisture and scouring force of dripping area. From the local native plant bank, suitable candidate species were selected based on environmental stress characteristic data of each zone; Based on the selected candidate species, the optimal seed ratio for each partition is obtained. Based on the optimal seed ratio and the microhabitat zoning model, differentiated precision sowing is carried out in different zones.

2. The method for intelligent configuration and restoration of native species populations in photovoltaic power plants according to claim 1, characterized in that, Environmental pressure characteristic data are quantified by constructing a three-dimensional environmental pressure characteristic vector, which includes cumulative photosynthetically active radiation, soil moisture availability index, and representative scour force.

3. The method for intelligent configuration and restoration of native species populations in photovoltaic power stations according to claim 1, characterized in that, Suitable candidate species were selected from the local native plant bank, including: For the shaded area, shade tolerance screening conditions are set, including: under shading conditions where the photosynthetically active radiation value is 20% of that in the full-light area, the net photosynthetic rate maintenance rate is not less than 60%, and the ratio of chlorophyll b to chlorophyll a is greater than 1.

2. For the dripping area, erosion resistance screening conditions are set, including: the root density is not lower than a preset first threshold, and the root tensile strength is not lower than a preset second threshold; For the full-sunlight zone, drought resistance screening conditions were set, including: when the soil moisture availability index is not greater than 0.25, the plant stomatal conductance is greater than 0.1 moles per square meter per second and the relative water content of the leaves is greater than 70%.

4. The method for intelligent configuration and restoration of native species populations in photovoltaic power stations according to claim 1, characterized in that, The optimal seed ratio for each partition is obtained by solving the problem, specifically: A multi-objective function is constructed with the optimization objectives of maximizing the weighted sum of expected vegetation survival rates in each zone, minimizing soil erosion across the entire station, and maximizing the functional diversity index of plant communities. Set optimization constraints that include a maximum total seeding amount, minimum coverage of a single species, and species compatibility. The multi-objective function is solved using a non-dominated sorting genetic algorithm to obtain a Pareto optimal solution set, and the solution with the highest comprehensive evaluation value is selected from the Pareto optimal solution set as the optimal seed ratio.

5. The method for intelligent configuration and restoration of native species populations in photovoltaic power stations according to claim 4, characterized in that, Minimizing soil erosion across the entire station is calculated by coupling the modified general soil loss equation RUSLE with root soil-fixing factors. The vegetation cover management factor is calculated using empirical formulas based on the weighted average of the root density of species in each zone.

6. The method for intelligent configuration and restoration of native species populations in photovoltaic power stations according to claim 1, characterized in that, Implementing differentiated precision seeding includes: The optimal seed ratio is associated with the digital vector map of the microhabitat zoning model to generate a work file containing zoning identifiers, spatial ranges, and corresponding seeding quantity lists; The seeding equipment, equipped with a positioning module and a multi-channel independent sowing device, automatically switches the corresponding seed storage bin or adjusts the sowing parameters based on real-time positioning information when the equipment enters different zones, according to the operation document, to ensure that the sowing amount per unit area is consistent with the optimal seed ratio.

7. The method for intelligent configuration and restoration of native species populations in photovoltaic power stations according to claim 1, characterized in that, Following the implementation of differentiated precision seeding, the process also includes post-implementation monitoring and feedback adjustments. At preset time points, the survival rate of vegetation in each zone was monitored to obtain the measured survival rate. The measured survival rate is compared with the expected survival rate predicted when solving the optimal seed ratio. If the deviation exceeds the preset threshold, the input data used for species screening or optimal ratio solving is corrected, and the screening and solving steps are re-executed to generate an updated seed formula.

8. The method for intelligent configuration and restoration of native species populations in photovoltaic power plants according to claim 7, characterized in that, Monitoring vegetation survival rate in each zone includes: obtaining measured data through ground quadrat surveys, calculating the Normalized Difference Vegetation Index (NDVI) by combining UAV multispectral remote sensing images, using machine learning models to invert and obtain the vegetation cover distribution map of the entire field, and extracting the measured cover and survival rate of each zone from it.

9. The method for intelligent configuration and restoration of native species populations in photovoltaic power plants according to claim 1, characterized in that, The maximum splash radius was determined based on hydrodynamic splashing experiments. Specifically, under laboratory conditions simulating preset extreme rainfall intensity, high-speed camera technology was used to determine the maximum splash distance of rainwater droplets on surface soil particles, and this distance was used as the basis for delineating the dripping zone.

10. A smart configuration and restoration system for native species populations in photovoltaic power plants, used to implement the method described in any one of claims 1 to 9, characterized in that, include: The data acquisition and partitioning module is used to obtain the geometric parameters of the photovoltaic array and, based on the geometric parameters, solar altitude angle, and maximum sputtering radius determined by hydrodynamic sputtering experiments, construct a digital model of microhabitat partitioning that includes shaded areas, dripping areas, and full-light areas. The environmental monitoring and quantification module is used to deploy a sensor network to collect and quantify environmental pressure characteristic data of each zone in real time. The environmental pressure characteristic data includes at least photosynthetically active radiation, soil moisture and drip scouring force. The species selection and optimization module has a built-in local native plant database, which is used to select suitable candidate species based on the environmental pressure characteristics data of each zone, and generate the optimal seed ratio for each zone by solving a multi-objective optimization problem. The precision seeding control module is used to receive the optimal seed ratio and the digital model of microhabitat zoning, generate operation files, and control the seeding equipment to carry out differentiated precision seeding in different zones based on real-time positioning information; The effect evaluation and feedback module is used to monitor the survival rate of vegetation after sowing and compare the measured data with the expected value. When the deviation exceeds the preset threshold, it triggers the correction of the input parameters of the species screening and optimization module.

Images