Optimization method for ecological restoration of mountain wind farm based on orthogonal test design

Abstract

The present application relates to the field of ecological restoration engineering, and more particularly to a mountain wind farm ecological restoration optimization method based on orthogonal test design. It comprises: determining the key factors including vegetation species, seeding density, substrate ratio and water-retaining agent dosage and the corresponding levels by obtaining altitude, slope and soil conditions, using the above information to generate a standardized orthogonal test table through orthogonal design, collecting soil improvement degree and vegetation survival rate indicators after implementation according to the table, and performing range and variance analysis on the indicators to quantify the primary and secondary order of the influence of each factor, and accordingly determining the optimal level combination to output the target restoration parameters. The present application can effectively avoid the high cost of traditional blind trial and error, improve the parameter optimization problem under the conflict of double indicators, and improve the pertinence and stability of mountain ecological restoration.

Description

Technical Field

[0001] This invention relates to the field of ecological restoration engineering, and in particular to an optimization method for ecological restoration of mountain wind farms based on orthogonal experimental design. Background Technology

[0002] With the rapid development of the mountain wind power industry, construction projects have severely disturbed the fragile mountain ecosystems. Mountain habitats are harsh, and their ecological restoration is a complex systems engineering project involving the coupling of multiple factors, including vegetation selection, seeding density, substrate ratio, and water-retaining agent dosage. To reconstruct stable ecological communities in arid and barren environments, it is essential to precisely control various construction parameters to jointly meet the core requirements of improving soil physical and chemical properties and achieving high vegetation survival rates.

[0003] Existing remediation solutions often rely on blind trial and error based on human experience or use undifferentiated parameter configurations. This results in high trial and error costs and unstable remediation effects. It is not only difficult to scientifically quantify the combined impact of multiple environmental factors, but also difficult to solve the problem of parameter optimization when soil improvement and vegetation survival are conflicting indicators. Summary of the Invention

[0004] To overcome the above shortcomings, this invention provides an optimization method for ecological restoration of mountain wind farms based on orthogonal experimental design, which aims to improve the problems of high cost and unstable results of traditional schemes that rely on blind trial and error, as well as the difficulty in finding optimal parameters under conflicting dual indicators.

[0005] This invention provides the following technical solution: an optimization method for ecological restoration of mountain wind farms based on orthogonal experimental design, comprising: S1. Obtain the altitude, slope and soil conditions of the mountain wind farm, and determine the key factors for ecological restoration accordingly. The key factors include vegetation type, seeding density, substrate ratio and water-retaining agent dosage, and set multiple levels for each key factor. S2. Using the key factors and the corresponding multiple levels, an orthogonal experimental design is carried out to construct an orthogonal experimental scheme containing multiple sets of repair parameter combinations, and a standardized orthogonal experimental table for experimental execution is generated. S3. Implement ecological restoration according to the multiple sets of restoration parameters in the orthogonal experimental table, and collect corresponding restoration evaluation indicators after restoration. The restoration evaluation indicators include the degree of soil improvement and the survival rate of vegetation. S4. Perform range analysis and variance analysis on the collected soil improvement degree and vegetation survival rate respectively to quantify and determine the order of importance of each key factor on the remediation evaluation index. S5. Based on the order of influence and the results of the range analysis and variance analysis, determine the optimal combination of levels for each key factor, obtain the target repair scheme, and output the standardized repair parameter combination.

[0006] Preferably, in step S1, the step of determining the key factors for ecological restoration includes: The altitude, slope, and soil conditions are converted into a digital gridded environmental parameter matrix; The gridded environmental parameter matrix is ​​correlated with a preset ecological degradation dataset to calculate the weight of each environmental factor's disturbance impact on vegetation damage and soil erosion. Based on the disturbance impact weights, collinear redundant variables are eliminated, and variables with values ​​greater than a preset threshold are extracted as core variables to complete the determination of the key factors for ecological restoration.

[0007] Preferably, in step S2, the step of performing orthogonal experimental design includes: Based on the number of the key factors and the order of the corresponding multiple levels, a matching target orthogonal base table is retrieved from a preset orthogonal table set; Using a matrix mapping operator, the key factors and the multiple levels are projected onto the rows and columns of the target orthogonal base table, so that the element arrangement has orthogonality and balance. Perform interaction checks and degree-of-freedom constraint verification between variables to complete the orthogonal experimental design under multi-factor, multi-level coupled conditions.

[0008] Preferably, in step S2, the step of generating a standardized orthogonal experimental table for experiment execution includes: The multidimensional matrix of the orthogonal experimental design is transformed into a set of process boundary parameters oriented towards the construction site, and a unique index identifier is configured. Based on the operational functional zoning characteristics of mountain wind farms, the matrix ratio and water-retaining agent dosage in the process boundary parameter set are converted into quantitative construction material parameters; The execution order of the construction batching parameters is adjusted according to the randomized block arrangement principle to eliminate environmental block errors, thereby generating the standardized orthogonal experimental table.

[0009] Preferably, in step S3, the step of collecting the corresponding repair evaluation indicators includes: Dynamically acquire vegetation canopy spectral data and soil physicochemical parameters within the restoration sample area; Based on the normalized vegetation index and growth density model, the vegetation canopy spectral data are calculated to generate the vegetation survival rate data covering the sample area; Data on soil bulk density and organic matter content in the sample plots were obtained, and the degree of soil improvement was calculated using a multi-index comprehensive evaluation model, thus completing the collection of the remediation evaluation indicators.

[0010] Preferably, in step S4, the steps of performing range analysis and variance analysis include: The average value of the repair evaluation index for each key factor in the standardized orthogonal experimental table at each specific level is calculated by category. Calculate the maximum absolute difference of the average value of the repair evaluation index at different levels of the same key factor, and perform the range analysis to characterize the effect span; Construct the sum of squared deviations derived from the average value of the repair evaluation index and calculate the F-test statistic to complete the analysis of variance.

[0011] Preferably, in step S4, the step of quantifying and determining the order of importance of the influence of each key factor on the repair evaluation index includes: Extract the absolute values ​​of the ranges corresponding to each key factor in the range analysis and sort them in descending order of numerical value. Obtain the significance probability values ​​corresponding to each key factor in the analysis of variance, and perform confidence testing and correction on the descending order ranking results; The sensitivity weights of each key factor are determined based on the revised arrangement, thereby quantifying and determining the order of influence.

[0012] Preferably, in step S5, the step of determining the preferred combination of levels for each key factor includes: Based on the effect curves of range analysis, the independent level levels corresponding to each key factor when a single repair evaluation index reaches its maximum value are determined. A multi-objective optimization algorithm is introduced, and a global trade-off function is established with the degree of soil improvement and the survival rate of vegetation as optimization axes to calculate the trade-off between different remediation evaluation indicators. Solve the Pareto non-dominated solution set of the global trade-off function, and map the target matching level under the dual-index synergistic state from it to determine the optimal combination of levels for each key factor.

[0013] Preferably, in step S5, the step of obtaining the target repair scheme and outputting the standardized repair parameter combination includes: The preferred level combination is input into a preset long-term succession prediction model to verify the stability of the preferred level combination under future extreme weather conditions. The optimal combination of levels that has passed stability verification is spatially superimposed with the obtained altitude, slope and soil conditions to generate the target restoration scheme that takes into account both vegetation community construction and micro-topography and soil and water conservation. The process variables in the target remediation scheme are converted into a digital control parameter matrix containing preset plant quantity, substrate thickness and water-retaining agent ratio, thus completing the output of the standardized remediation parameter combination.

[0014] The present invention has the following beneficial effects: 1. In this invention, orthogonal experimental design and Pareto multi-objective optimization algorithm are introduced into the field of mountain ecological restoration. By scientifically combining and globally weighing multiple factors and levels such as vegetation type and substrate ratio, the high cost and drawbacks of traditional blind trial and error are avoided, and parameter calculation under the conflict between soil improvement and vegetation survival is realized.

[0015] 2. In this invention, a closed-loop evaluation system is constructed that integrates quantitative modeling of multi-source data from air and ground with rigorous statistical analysis of range and variance. By using UAV spectral extraction of vegetation indices and combining them with measured soil parameters, along with a randomized block arrangement algorithm, the subjective errors of complex mountain micro-topography changes and traditional manual observations are effectively avoided, and the core control variables are locked.

[0016] 3. In this invention, a long-cycle succession prediction model is pre-set to verify the noise resistance stability of the preferred scheme under extreme weather conditions, and the macroscopic theoretical variables are transformed into a digital control parameter matrix containing process dimensions such as preset plant quantity and substrate thickness, thereby realizing the transformation of the theoretical optimization model into on-site automated construction instructions. Attached Figure Description

[0017] Figure 1 This is a flowchart of the ecological restoration optimization method for mountain wind farms based on orthogonal experimental design proposed in this invention; Figure 2 This is a flowchart of the quantitative dimensionality reduction process for environmental factors based on multi-source data and correlation analysis proposed in this invention. Figure 3 This is a flowchart illustrating the orthogonal experimental design and construction transformation of the coupled randomized block arrangement proposed in this invention. Figure 4 This is a flowchart of the multi-source data quantification modeling and objective evaluation process for air-ground collaboration proposed in this invention. Figure 5 This is a diagram of the dual-channel statistical mechanism for range and variance that integrates confidence verification correction proposed in this invention; Figure 6 This is a prediction graph of Pareto optimization and DNN long-term stability for dual-index conflict proposed in this invention. Detailed Implementation

[0018] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] In embodiments of the present invention, the present invention provides an optimization method for ecological restoration of mountain wind farms based on orthogonal experimental design, such as... Figure 1 As shown, it includes: S1. Obtain the altitude, slope and soil conditions of the mountain wind farm, and determine the key factors for ecological restoration based on this. The key factors include vegetation type, seeding density, substrate ratio and water-retaining agent dosage, and set multiple levels for each key factor. Furthermore, step S1, the step of identifying key factors for ecological restoration, includes: Elevation, slope, and soil conditions are converted into a digital gridded environmental parameter matrix; The gridded environmental parameter matrix is ​​correlated with the preset ecological degradation dataset to calculate the weight of each environmental factor's disturbance impact on vegetation damage and soil erosion. Based on the weighting of disturbance effects, collinear redundant variables are eliminated, and variables with values ​​greater than a preset threshold are extracted as core variables to determine the key factors for ecological restoration.

[0020] Specifically, the environmental baseline data for the mountain wind farm is obtained through the following steps: A digital elevation model of the area to be restored is acquired using oblique photogrammetry by unmanned aerial vehicles (UAVs), and elevation and slope data are extracted using spatial analysis tools. Soil samples covering the area are collected and subjected to physicochemical analysis to obtain soil bulk density and organic matter content as soil condition data. The entire area to be restored is divided into a two-dimensional grid of the same size, specifically set at 10 meters by 10 meters. The elevation, slope, and soil condition values ​​corresponding to the center point of each grid are extracted. The data from all grids are integrated according to their spatial location to construct a digital gridded environmental parameter matrix. Let the total number of grids be... The total number of environmental parameter categories is The constructed gridded environment parameter matrix is , for OK A real matrix of columns, where the matrix elements Indicates the first The corresponding grid number is the first The specific measured values ​​for each environmental parameter.

[0021] A pre-defined ecological degradation dataset covering areas disturbed by historical wind turbine construction was retrieved. This dataset contains vegetation damage rates and soil erosion moduli that correspond one-to-one with the spatial locations of the aforementioned two-dimensional grid. A correlation analysis was performed between the constructed digitized gridded environmental parameter matrix and the pre-defined ecological degradation dataset under spatial location matching. The disturbance impact weights of each environmental factor on the two degradation indicators of vegetation damage and soil erosion were calculated, respectively. The calculation formulas are as follows: ; In the above formula, Indicates the first The environmental factor affects the first The weight of the disturbance effect on the degradation index. The value is 1 or 2, representing the vegetation damage rate and the soil erosion modulus, respectively. For the first The first grid A number of environmental parameter values. For the first in all grids The average value of each environmental parameter. For the preset ecological degradation dataset, the first The corresponding grid number is the first Historical actual observation values ​​of the degradation index. For the first in all grids The average value of each degradation index.

[0022] Based on the calculated perturbation impact weights, a multivariate linear regression model was constructed to test the collinearity among the environmental parameter variables. For variable combinations exhibiting collinearity redundancy (i.e., the set of associated variables with a calculated variance inflation factor greater than 10), a screening process was performed according to their weights, directly removing variables with smaller perturbation impact weights and retaining those with the largest perturbation impact weights. A threshold filter was then applied to the remaining variables after removing collinearity redundancy, extracting variables with perturbation impact weights greater than a preset threshold of 0.5 as core variables. Based on the extracted core variables, key ecological restoration factors requiring human intervention were determined. Topographic core variables such as altitude and slope corresponded to two key factors: vegetation type and seeding density. Geological core variables such as soil bulk density and organic matter content corresponded to two key factors: substrate ratio and water-retaining agent dosage.

[0023] For the four identified key ecological restoration factors, and considering the extreme weather and barren geological characteristics of the mountain wind farm, three discrete restoration process levels were set to form the basic parameter space for the orthogonal experiment. For the vegetation type factor, three levels were set: shallow-rooted herbaceous plants, deep-rooted shrubs, and a mixture of shrubs and grasses. For the seeding density factor, three levels were set: 10 grams per square meter, 15 grams per square meter, and 20 grams per square meter. For the substrate ratio factor, three levels were set: the mass ratio of topsoil to specialized bio-organic fertilizer was 3:1, 5:1, and 7:1. For the water-retaining agent dosage factor, three levels were set: 5 grams per square meter, 10 grams per square meter, and 15 grams per square meter. The parameter settings for each factor and its corresponding levels were completed, providing the preliminary data boundary for generating a standardized orthogonal experimental table for the experiment.

[0024] This step achieves quantitative dimensionality reduction of environmental factors through data analysis, avoiding the blindness of subjectively setting restoration factors and ensuring that the extracted experimental parameters are highly compatible with the actual ecological characteristics of the wind farm.

[0025] S2. Utilize key factors and their corresponding multiple levels to design orthogonal experiments, construct orthogonal experimental schemes containing multiple sets of repair parameter combinations, and generate standardized orthogonal experimental tables for experiment execution; Furthermore, step S2, the steps of conducting orthogonal experimental design, include: Based on the number of key factors and the order of the corresponding multiple levels, the matching target orthogonal base table is retrieved from the preset orthogonal table set; By using matrix mapping operators, key factors and multiple levels are projected onto the rows and columns of the target orthogonal base table, so that the element arrangement has orthogonality and balance. Perform interaction checks and degree-of-freedom constraint verification between variables to complete orthogonal experimental design under multi-factor, multi-level coupled conditions.

[0026] Further, step S2, the step of generating a standardized orthogonal experimental table for experiment execution, includes: The multidimensional matrix of the orthogonal experimental design is transformed into a set of process boundary parameters oriented towards the construction site, and a unique index identifier is configured. Based on the operational functional zoning characteristics of mountain wind farms, the matrix ratio and water-retaining agent dosage in the process boundary parameter set are transformed into quantitative construction material parameters; The execution order of construction batching parameters is adjusted according to the randomized block arrangement principle to eliminate environmental block errors and generate a standardized orthogonal experimental table.

[0027] Specifically, based on the four key factors determined in the preceding steps—vegetation type, sowing density, substrate ratio, and water-retaining agent dosage—and the three specific process levels set for each factor, a search is performed in a pre-defined orthogonal array database to obtain a target orthogonal base table that can accommodate the four-factor, three-level structure. A 9-row, 4-column orthogonal array is specifically selected. The target orthogonal base table is set as a matrix. , The number of rows is the total number of trials. , The number of columns represents the total number of key factors. .matrix elements in Indicates the first The key factor in The standard level designation corresponding to this experiment is used. A matrix mapping operator is introduced to transform the standard level designation into actual engineering parameters, constructing the mapped process matrix. Process matrix elements in Indicates the first The first group of repair parameter combinations The specific engineering values ​​for each key factor.

[0028] After parameter projection is completed, the system performs interaction checks between variables and degree-of-freedom constraint verification. The level order of each key factor is extracted. The minimum degrees of freedom required for the orthogonal experiment are calculated. The formula for verifying the degree of freedom constraints is as follows: ; In the above formula, This represents the total degrees of freedom of the orthogonal experimental design, and its value is equal to the total number of trials. Subtract 1. The total number of key factors. Representing the The level order of each key factor is set. This represents the degrees of freedom for the interaction between factors. Since this stage focuses on screening the main effects of independent key factors, the degrees of freedom for the interaction are set to 0. Substituting the known data, the total degrees of freedom is 8, and the sum of the degrees of freedom of all factors is also 8. This satisfies the constraint that the total degrees of freedom are not less than the sum of the degrees of freedom of each factor, confirming the validity of the orthogonal experimental design under multi-factor, multi-level coupling.

[0029] Extraction process matrix Each row in the dataset serves as an independent set of multi-dimensional process boundary parameters, and each set is assigned a unique index identifier generated by concatenating timestamps and feature codes. Based on the operational functional zoning characteristics of mountain wind farms, the specific operational area values ​​of the sample plot units to be restored are obtained. and the design substrate thickness values Using the aforementioned area and thickness values, the matrix proportions and water-retaining agent dosages, which are in a relatively proportional state within the set of process boundary parameters, are converted into quantitative construction material parameters that directly guide construction. Let the average bulk density of the matrix mixture be... The mass ratio of topsoil to special bio-organic fertilizer in the substrate mix is: The required quantity of specialized bio-organic fertilizer for a single sample plot unit and quality of imported soil The calculation formula is as follows: ; ; In the above formula, This indicates the mass of specialized bio-organic fertilizer required for each sample plot unit, expressed in kilograms. This indicates the required mass of topsoil for a sample plot unit, expressed in kilograms. This indicates the work area, in square meters. This indicates the thickness of the substrate to be laid, in meters. This indicates the average bulk density of the matrix mixture, expressed in kilograms per cubic meter. This is a dimensionless mass ratio. Simultaneously, multiply the amount of water-retaining agent in the process parameters by the working area. The total mass of water-retaining agent required for a single plot unit was obtained.

[0030] After obtaining nine test schemes with quantitative construction material parameters, a randomized block arrangement algorithm was introduced to adjust the execution order of the construction material parameters. A non-repeating pseudo-random integer sequence of length 9 was generated and used as the basis for spatial allocation mapping. The nine sets of repair parameter combinations were randomly assigned to nine adjacent test grid blocks in the wind farm. By shuffling the linear correspondence between parameter combinations and spatial locations, environmental block errors caused by gradual changes in mountain slope or illumination angle were eliminated. Finally, a standardized orthogonal test table with a unique index identifier, quantitative material details, and random spatial coordinates was output.

[0031] This step enables a seamless transformation from theoretical experimental design to actual engineering construction material parameters, ensuring scientific balance in the experiment while eliminating error interference caused by gradual changes in terrain and environment.

[0032] S3. Implement ecological restoration according to the combination of multiple restoration parameters in the orthogonal experimental table, and collect the corresponding restoration evaluation indicators after restoration. The restoration evaluation indicators include the degree of soil improvement and the survival rate of vegetation. Furthermore, step S3, the step of collecting the corresponding remediation evaluation indicators, includes: Dynamically acquire vegetation canopy spectral data and soil physicochemical parameters within the restoration sample area; Based on the normalized vegetation index and growth density model, vegetation canopy spectral data are calculated to generate vegetation survival rate data for the covered sample plot area. Data on soil bulk density and organic matter content in the sample plots were obtained, and the degree of soil improvement was calculated using a multi-index comprehensive evaluation model, thus completing the collection of remediation evaluation indicators.

[0033] Specifically, following standardized orthogonal experimental design parameters including vegetation species, sowing density, substrate ratio, and water-retaining agent dosage, substrate laying and vegetation sowing were completed in each experimental grid block corresponding to the mountain wind farm. After the ecological restoration work was completed and one full plant growing season had passed, a dynamic data acquisition program for restoration evaluation indicators was initiated. A drone equipped with a multispectral camera flew over the restoration sample plot along a set route to acquire multispectral image data of the vegetation canopy covering all grid blocks, extracting near-infrared and red reflectance as plant spectral characteristic data. Simultaneously, soil samples from each grid block were collected using the standard ring cutter method and the potassium dichromate volumetric method. Laboratory physicochemical measurements were then performed to obtain the actual bulk density and organic matter content of the soil in each block.

[0034] Based on the acquired multispectral image data of the vegetation canopy, the vegetation survival rate of each experimental block was calculated using the normalized vegetation index and growth density model. First, the near-infrared and red reflectance of each pixel in the image were extracted and substituted into the normalized vegetation index formula for basic calculation.

[0035] ; In the formula, The normalized vegetation index value represents a single pixel. The near-infrared reflectance of the vegetation canopy The red light reflectance represents the vegetation canopy. After calculating the normalized vegetation index for each pixel, a growth density model is constructed using the pixel-based binary search method to calculate the actual vegetation cover density within a single grid block.

[0036] ; In the formula, This represents the actual vegetation cover density of a single grid block. This is the average arithmetic value of the normalized vegetation index for all pixels within this block. This represents the empirical baseline value of the normalized vegetation index for bare soil pixels in the mountainous wind farm area. This represents the empirical upper limit of the normalized vegetation index for pixels with pure vegetation cover in this area. Combining the initial seeding density parameters set in the orthogonal experimental table with the theoretical maximum cover area of ​​the corresponding vegetation species, the standard theoretical cover density for the current growing season is derived. Dividing the actual vegetation cover density by the standard theoretical cover density finally generates quantified vegetation survival rate data for the covered sample plot area.

[0037] Soil bulk density and organic matter content data from the completed sample plots were obtained, and the degree of soil improvement was calculated using a multi-index comprehensive evaluation model. Given that soil bulk density is an inverse physical indicator and organic matter content is a positive chemical indicator, the range standardization method was used to transform the two original physicochemical measurement data into a dimensionless evaluation index.

[0038] ; ; In the formula, Represents the standardized index of soil bulk density. This represents the measured soil bulk density value of the sample plot. and These represent the maximum bulk density of the extremely degraded soil in the wind farm that has not been repaired and the minimum bulk density of the original, undamaged soil, respectively. Represents the organic matter standardization index. This represents the measured soil organic matter content of the sample plot. and These represent the maximum organic matter content in pristine, undamaged soil and the minimum organic matter content in severely degraded soil, respectively. After variable standardization, a multi-index comprehensive evaluation model is used to calculate the final degree of soil improvement.

[0039] ; In the formula, The numerical value representing the degree of soil improvement This represents the perturbation weight values ​​of the soil bulk density index, which are pre-assigned using the analytic hierarchy process (AHP). The value represents the perturbation weight of the pre-assigned organic matter content index, and the sum of the two weight values ​​equals 1.

[0040] This step, through an air-ground collaborative data acquisition mechanism and a multi-source data quantification modeling algorithm, transforms the on-site physical and chemical remediation feedback characteristics into standardized mathematical indicators. This effectively eliminates the subjective assessment errors of traditional manual observation and lays an objective data foundation for the subsequent extraction of the optimal combination of remediation parameters.

[0041] S4. Range analysis and variance analysis were performed on the collected soil improvement degree and vegetation survival rate to quantify and determine the order of importance of the influence of each key factor on the remediation evaluation index. Furthermore, step S4, which involves performing range analysis and variance analysis, includes: The average value of the repair evaluation index for each key factor in the standardized orthogonal experimental table at each specific level was calculated by classification. Calculate the maximum absolute difference of the average value of the repair evaluation index at different levels of the same key factor, and perform range analysis to characterize the effect span; Construct the sum of squared deviations derived from the average value of the repair evaluation index and calculate the F-test statistic to complete the analysis of variance.

[0042] Furthermore, step S4, which quantifies and determines the order of importance of the impact of each key factor on the remediation evaluation index, includes: Extract the absolute values ​​of the ranges corresponding to each key factor in the range analysis and sort them in descending order of numerical value. Obtain the significance probability values ​​of each key factor in the analysis of variance, and perform confidence tests and corrections on the results of the descending sorting; Based on the revised arrangement, the sensitivity weights of each key factor are determined, thereby quantifying and determining the order of influence.

[0043] Specifically, soil improvement and vegetation survival rate data were obtained after the completion of all nine groups of experiments in the standardized orthogonal experimental table. Independent statistical analyses were conducted for these two remediation evaluation indicators. Taking a single remediation evaluation indicator as an example, the average remediation evaluation indicator for each key factor at each specific level was calculated. The total number of key factors was set to [number missing]. The total number of levels is The number of repeated trials at a single level is The calculation formula is as follows: ; In the above formula, Indicates the first The key factor in The average value of the repair evaluation index at a specific level. Indicates the first The key factor is taken as the first The first level The actual calculated values ​​of soil improvement degree or vegetation survival rate corresponding to this experiment. Based on the calculated average values, the maximum absolute difference of the average values ​​of the remediation evaluation indicators at different levels of the same key factor was calculated to perform range analysis. The calculation formula is as follows: ; In the above formula, Indicates the first The absolute values ​​of the ranges corresponding to key factors. This value intuitively depicts the comprehensive effect span of the final ecological restoration of the wind farm when engineering parameters such as matrix ratio or water-retaining agent dosage vary within their set upper and lower limits. After completing the range calculation, the sum of squared deviations derived from the average value of the restoration evaluation indicators is constructed, and the test statistic is calculated to perform analysis of variance. The calculation formula is as follows: ; ; In the above formula, Indicates the first The sum of squared factor deviations for each key factor. This represents the overall average value of the repair evaluation index across all 9 groups of trials. Indicates the first The variance test statistics for each key factor. This represents the sum of squared error deviations corresponding to the total error of the orthogonal experiment. and They represent the first The theoretical degrees of freedom corresponding to each key factor and the theoretical degrees of freedom corresponding to experimental errors.

[0044] The absolute values ​​of the ranges for four key factors—vegetation type, sowing density, substrate ratio, and water-retaining agent dosage—were extracted from the range analysis and initially sorted in descending order of value. The significance probability values ​​corresponding to the test statistics of each key factor in the variance analysis were obtained and compared with a preset statistical confidence threshold of 0.05. Confidence tests and logical corrections were performed on the initial descending sort results. If the significance probability value of a currently ranked key factor is greater than 0.05, indicating that its range fluctuation may originate from environmental errors, its ranking in the queue was adjusted to after key factors with a significance probability value less than 0.05. Based on the corrected final sequence sorting results, higher sensitivity weight values ​​were assigned to the higher-ranked key factors, and lower sensitivity weight values ​​were assigned to the lower-ranked key factors, thereby quantifying and ultimately determining the order of importance of the influence of each construction key factor on the degree of soil improvement and vegetation survival rate.

[0045] This step, by combining range calculation with a rigorous variance test algorithm, effectively eliminates the interference of random environmental errors in the complex mountainous field tests of wind farms, and objectively locks in the core control variables that determine the success or failure of ecological restoration in the region.

[0046] S5. Based on the order of influence and the results of range analysis and variance analysis, determine the optimal combination of levels for each key factor, obtain the target repair scheme, and output the standardized repair parameter combination.

[0047] Furthermore, step S5, which involves determining the optimal combination of levels for each key factor, includes: Based on the effect curves of range analysis, the independent level levels corresponding to each key factor when a single repair evaluation index reaches its maximum value are determined. A multi-objective optimization algorithm is introduced, and a global trade-off function is established with soil improvement degree and vegetation survival rate as optimization axes to calculate the trade-off between different remediation evaluation indicators. Solve the Pareto non-dominated solution set of the global trade-off function, and map the target matching level under the dual-index synergistic state to determine the optimal combination of levels for each key factor.

[0048] Further, in step S5, the step of obtaining the target repair scheme and outputting the standardized repair parameter combination includes: The preferred level combination is input into a preset long-term succession prediction model to verify the stability of the preferred level combination under future extreme weather conditions. The optimal combination of levels that has passed stability verification is spatially superimposed with the obtained altitude, slope and soil conditions to generate the target restoration scheme that takes into account both vegetation community construction and micro-topography and soil and water conservation. The process variables in the target remediation scheme are converted into a digital control parameter matrix containing preset plant quantity, substrate thickness and water-retaining agent ratio, thus completing the output of the standardized remediation parameter combination.

[0049] Specifically, the average values ​​of the remediation evaluation indicators for each key factor output from the range analysis were extracted at different levels. Using the level level as the horizontal axis and the remediation evaluation indicator value as the vertical axis, soil improvement effect curves and vegetation survival rate effect curves were plotted respectively. A peak-finding algorithm was used to locate the independent level level corresponding to each key factor when a single remediation evaluation indicator reaches its maximum value. For process parameters where the independent level levels corresponding to the maximum soil improvement value and the maximum vegetation survival rate value conflict, a multi-objective optimization algorithm was introduced to establish a global trade-off function. The objective optimization input parameters were set as a combination vector including vegetation type, seeding density, substrate ratio, and water-retaining agent dosage. The following mathematical metric formula for the global trade-off function was established.

[0050] ; In the formula, The distance function value represents the synergistic trade-off between the two indicators. This represents the soil improvement level calculated after inputting a specific combination of remediation parameters. This represents the vegetation survival rate calculated after inputting this combination. The theoretical maximum value of the degree of soil improvement in the representative effect curve. The theoretical maximum value of vegetation survival rate in the representative effect curve. and These represent the trade-off preference coefficients for soil improvement and vegetation survival, respectively, and their sum is strictly 1. By traversing all orthogonal experimental level combinations, the Pareto non-dominated solution set of this global trade-off function is obtained, i.e., the set that makes the distance function... The eigenvector with the smallest numerical value. The target matching level under the dual-index synergistic state is mapped from this solution set, thereby determining the optimal combination of levels for each key factor.

[0051] The determined optimal level combination is input into a pre-defined long-term succession prediction model. This prediction model employs a DNN architecture, specifically including one input layer, four fully connected hidden layers, and one output layer to ensure the complete reproducibility of the technical solution. By injecting 10 years of historical extreme meteorological drought and rainstorm time-series data from the Northwest mountainous region into the model's input layer, the vegetation degradation rate of the optimal level combination in future succession cycles is calculated. If the vegetation degradation rate is less than 15%, the combination is deemed to have the ability to withstand extreme weather events and passes stability verification. The process values ​​corresponding to the optimal level combination that passed stability verification are spatially overlaid with the actual altitude, slope, and soil conditions of the wind farm in a geographic information system to generate a target restoration scheme that balances vegetation community construction and micro-topography soil and water conservation. The macroscopic process variables in the target restoration scheme are quantified and extracted, converting them into a digital control parameter matrix for mechanized construction. The standardized control parameter matrix of the output is set as follows. .

[0052] ; In the formula, This represents a digital control parameter matrix for a single repair grid block. This represents the preset absolute amount of plants to be added, calculated based on the optimal sowing density and vegetation type, in kilograms per square meter. The thickness of the substrate, measured in centimeters, represents the substrate thickness determined based on the optimal substrate ratio and terrain slope. This represents the absolute mass ratio of water-retaining agent based on the optimal dosage and target soil conditions. The standardized remediation parameter matrix is ​​then assembled and output, and directly configured to the electronic control terminal of the on-site construction equipment.

[0053] This step, through multi-objective mathematical optimization and calculation verification of a clear network architecture, solves the problem of selecting process parameters that conflict with each other among different ecological evaluation indicators, and finally outputs standardized construction data with high noise resistance and stability that can be directly used to guide on-site mechanized operations.

[0054] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An optimization method for ecological restoration of mountain wind farms based on orthogonal experimental design, characterized in that, include: S1. Obtain the altitude, slope and soil conditions of the mountain wind farm, and determine the key factors for ecological restoration accordingly. The key factors include vegetation type, seeding density, substrate ratio and water-retaining agent dosage, and set multiple levels for each key factor. S2. Using the key factors and the corresponding multiple levels, an orthogonal experimental design is carried out to construct an orthogonal experimental scheme containing multiple sets of repair parameter combinations, and a standardized orthogonal experimental table for experimental execution is generated. S3. Implement ecological restoration according to the multiple sets of restoration parameters in the orthogonal experimental table, and collect corresponding restoration evaluation indicators after restoration. The restoration evaluation indicators include the degree of soil improvement and the survival rate of vegetation. S4. Perform range analysis and variance analysis on the collected soil improvement degree and vegetation survival rate respectively to quantify and determine the order of importance of each key factor on the remediation evaluation index. S5. Based on the order of influence and the results of the range analysis and variance analysis, determine the optimal combination of levels for each key factor, obtain the target repair scheme, and output the standardized repair parameter combination.

2. The optimization method for ecological restoration of mountain wind farms based on orthogonal experimental design according to claim 1, characterized in that, Step S1, the step of determining the key factors for ecological restoration, includes: The altitude, slope, and soil conditions are converted into a digital gridded environmental parameter matrix; The gridded environmental parameter matrix is ​​correlated with a preset ecological degradation dataset to calculate the weight of each environmental factor's disturbance impact on vegetation damage and soil erosion. Based on the disturbance impact weights, collinear redundant variables are eliminated, and variables with values ​​greater than a preset threshold are extracted as core variables to complete the determination of the key factors for ecological restoration.

3. The optimization method for ecological restoration of mountain wind farms based on orthogonal experimental design according to claim 1, characterized in that, Step S2, the step of performing orthogonal experimental design, includes: Based on the number of the key factors and the order of the corresponding multiple levels, a matching target orthogonal base table is retrieved from a preset orthogonal table set; Using a matrix mapping operator, the key factors and the multiple levels are projected onto the rows and columns of the target orthogonal base table, so that the element arrangement has orthogonality and balance. Perform interaction checks and degree-of-freedom constraint verification between variables to complete the orthogonal experimental design under multi-factor, multi-level coupled conditions.

4. The method for ecological restoration optimization of mountain wind farms based on orthogonal experimental design according to claim 1, characterized in that, Step S2, the step of generating a standardized orthogonal experimental table for experiment execution, includes: The multidimensional matrix of the orthogonal experimental design is transformed into a set of process boundary parameters oriented towards the construction site, and a unique index identifier is configured. Based on the operational functional zoning characteristics of mountain wind farms, the matrix ratio and water-retaining agent dosage in the process boundary parameter set are converted into quantitative construction material parameters; The execution order of the construction batching parameters is adjusted according to the randomized block arrangement principle to eliminate environmental block errors, thereby generating the standardized orthogonal experimental table.

5. The optimization method for ecological restoration of mountain wind farms based on orthogonal experimental design according to claim 1, characterized in that, Step S3, the step of collecting the corresponding repair evaluation indicators, includes: Dynamically acquire vegetation canopy spectral data and soil physicochemical parameters within the remediation sample area; Based on the normalized vegetation index and growth density model, the vegetation canopy spectral data are calculated to generate the vegetation survival rate data covering the sample area; Data on soil bulk density and organic matter content in the sample plots were obtained, and the degree of soil improvement was calculated using a multi-index comprehensive evaluation model, thus completing the collection of the remediation evaluation indicators.

6. The optimization method for ecological restoration of mountain wind farms based on orthogonal experimental design according to claim 1, characterized in that, Step S4, the steps of performing range analysis and variance analysis, include: The average value of the repair evaluation index for each key factor in the standardized orthogonal experimental table at each specific level is calculated by category. Calculate the maximum absolute difference of the average value of the repair evaluation index at different levels of the same key factor, and perform the range analysis to characterize the effect span; Construct the sum of squared deviations derived from the average value of the repair evaluation index and calculate the F-test statistic to complete the analysis of variance.

7. The optimization method for ecological restoration of mountain wind farms based on orthogonal experimental design according to claim 1, characterized in that, Step S4, the step of quantifying and determining the order of importance of the influence of each key factor on the repair evaluation index, includes: Extract the absolute values ​​of the ranges corresponding to each key factor in the range analysis and sort them in descending order of numerical value. Obtain the significance probability values ​​corresponding to each key factor in the analysis of variance, and perform confidence testing and correction on the descending order ranking results; The sensitivity weights of each key factor are determined based on the revised arrangement, thereby quantifying and determining the order of influence.

8. The method for ecological restoration optimization of mountain wind farms based on orthogonal experimental design according to claim 1, characterized in that, Step S5, the step of determining the preferred combination of levels for each key factor, includes: Based on the effect curves of range analysis, the independent level levels corresponding to each key factor when a single repair evaluation index reaches its maximum value are determined. A multi-objective optimization algorithm is introduced, and a global trade-off function is established with the degree of soil improvement and the survival rate of vegetation as optimization axes to calculate the trade-off between different remediation evaluation indicators. Solve the Pareto non-dominated solution set of the global trade-off function, and map the target matching level under the dual-index synergistic state from it to determine the optimal combination of levels for each key factor.

9. The optimization method for ecological restoration of mountain wind farms based on orthogonal experimental design according to claim 1, characterized in that, Step S5, the step of obtaining the target repair scheme and outputting the standardized repair parameter combination, includes: The preferred level combination is input into a preset long-term succession prediction model to verify the stability of the preferred level combination under future extreme weather conditions. The optimal combination of levels that has passed stability verification is spatially superimposed with the obtained altitude, slope and soil conditions to generate the target restoration scheme that takes into account both vegetation community construction and micro-topography and soil and water conservation. The process variables in the target remediation scheme are converted into a digital control parameter matrix containing preset plant quantity, substrate thickness and water-retaining agent ratio, thus completing the output of the standardized remediation parameter combination.

Images