CN-121980633-A - Wind power installation potential evaluation method based on marginal constraint model

CN121980633ACN 121980633 ACN121980633 ACN 121980633ACN-121980633-A

Abstract

The invention provides a wind power installation potential evaluation method based on a marginal constraint model, and belongs to the field of maximum developable installation scale evaluation of wind power plants. The method comprises the steps of constructing a feature vector of a grid corresponding to a wind power plant site selection candidate area, inputting a random forest model to obtain a suitability probability prediction value of the grid for a wind power development category, screening a final candidate area, obtaining a theoretical maximum installed density of the final candidate area through a preset constraint line equation reflecting the theoretical maximum installed density of the suitability probability based on the suitability probability prediction value of the final candidate area, and multiplying the installed density by a utilization coefficient of a land utilization type and gradient grade combination corresponding to the final candidate area to obtain an installed potential evaluation result of the final candidate area. The method creatively combines the constraint line model and the random forest algorithm, provides a more accurate and efficient mode for evaluating the installation potential of the wind power plant, and aims to provide decision support for wind energy development.

Inventors

YIN YUNHE
ZHANG XUFANG

Assignees

中国科学院地理科学与资源研究所

Dates

Publication Date: 20260505
Application Date: 20251211

Claims (8)

1. The wind power installation potential assessment method based on the marginal constraint model is characterized by comprising the following steps of: constructing a feature vector of a grid corresponding to the wind power plant site selection candidate area; Inputting the feature vector into a preset random forest model to obtain a suitability probability prediction value of wind power development of a grid corresponding to the wind power plant site selection candidate region, and further screening out a final candidate region; Obtaining the theoretical maximum installed density of the final candidate region through a preset constraint line equation reflecting the theoretical maximum installed density corresponding to the suitability probability based on the suitability probability predicted value of the grid corresponding to the final candidate region; Acquiring a utilization coefficient of a land utilization type and gradient grade combination corresponding to the final candidate area; and multiplying the theoretical maximum installed density of the final candidate region by the utilization coefficient to obtain an installed potential evaluation result of the final candidate region.
2. The method of claim 1, wherein constructing feature vectors of a grid corresponding to a wind farm site selection candidate region comprises: 1) Collecting multi-source grid data and vector data of a wind power plant site selection candidate area, and unifying a space coordinate system; 2) Based on the result of the step 1), calculating the characteristic index of the grid corresponding to the wind power plant site selection candidate area, wherein the characteristic index comprises the following steps: 2-1) calculating wind energy potential indexes, wherein the wind energy potential indexes comprise a wind power density index and a wind resource stability index; wherein, wind power density index calculation expression is as follows: Wherein, the Is wind power density, Is air density, Time-by-time wind speed of 100 meters in height; The calculation expression of the wind resource stability index is as follows: Wherein, the Is an index of the stability of the wind resource, For an event with a wind speed below 3 m/s or above 25m/s to occur at a time, As the number of hours in a year, Is the total number of events occurring in a year with a wind speed below 3 m/s or above 25 m/s; 2-2) calculating socioeconomic indexes including GDP, distance from road, distance from power grid; 2-3) calculating geographic environment indexes including gradient, elevation and extreme temperature frequency; 2-4) calculating disaster risk level indexes, wherein the disaster risk level indexes comprise geological disaster indexes and meteorological disaster indexes; 2-5) calculating an ecological safety index, wherein the ecological safety index comprises a proximity score, a vegetation coverage score and a biodiversity score, and the method comprises the following specific steps of: 2-5-1) calculating a proximity score comprising: Determining grids corresponding to the ecological source land and the ecological corridor, setting 10 kilometers away from the ecological corridor or the ecological source land as the maximum influence distance, and calculating the proximity scores of the grids corresponding to the sample, the ecological source land and the ecological corridor by using a negative exponential function, wherein the expression is as follows: In the formula, In order to calculate the distance value to be obtained, Is the attenuation coefficient; 2-5-2) vegetation coverage score; obtaining a value of a grid corresponding to each sample in the vegetation coverage grid data product, namely a vegetation coverage score of the sample, and directly taking the value as a vegetation coverage characteristic value of the sample; 2-5-3) calculating a biodiversity score; Acquiring a value of a grid corresponding to each sample in the biodiversity index spatial distribution data product, namely biodiversity score of the sample, and directly taking the value as biodiversity characteristic value of the sample; 3) And respectively normalizing index values within a range of non-0-1 in the characteristic indexes of each grid, converting the index values into characteristic values, and forming the characteristic vector of each grid by all the characteristic values of each grid.
3. The method of claim 2, wherein prior to said inputting the feature vector into a pre-set random forest model, the method further comprises: training the random forest model; said training said random forest model comprising: 1) Constructing a sample set consisting of positive samples and negative samples, wherein the positive samples are grids corresponding to the built wind power plant, the negative samples are grids not building the wind power plant, and the number of the positive samples is the same as that of the negative samples; Setting the label value corresponding to the positive sample to be 1 to indicate that the wind power plant is suitable to be built, and setting the label value corresponding to the negative sample to be 0 to indicate that the wind power plant is not suitable to be built; 2) Collecting multi-source raster data and vector data of grids corresponding to each sample in a sample set, and unifying a space coordinate system; 3) Calculating the characteristic index of the grid corresponding to each sample in the sample set based on the result of the step 2); 4) Respectively normalizing index values within a range of non-0-1 in the characteristic indexes of the grids corresponding to each sample, converting the index values into characteristic values, and forming the characteristic vector of each sample by all the characteristic values of the grids corresponding to each sample; 5) Forming a data sample set by the feature vector of each sample and the corresponding label, and then randomly dividing the data sample set into a training set and a testing set according to a preset proportion; 6) And constructing a random forest model, training the model by using the training set, and checking the trained model by using the testing set to obtain a final random forest model.
4. A method according to claim 3, further comprising: And generating a grid diagram of the wind power plant site selection suitability probability distribution based on the suitability probability prediction value of wind power development performed by the grid corresponding to the wind power plant site selection candidate region output by the random forest model.
5. The method as recited in claim 4, further comprising: And obtaining the feature vector of the grid corresponding to the built wind power plant, inputting the feature vector into the random forest model, then outputting a suitability probability predicted value corresponding to the built wind power plant from the random forest model, selecting the minimum value as a candidate region screening threshold, and taking the wind power plant site candidate region as a final candidate region if the suitability probability predicted value of any wind power plant site candidate region is not smaller than the threshold.
6. The method as recited in claim 5, further comprising: And generating a scatter diagram reflecting the relation between the suitability probability and the installed density based on the data of the established wind power plants by using the grid diagram of the wind power plant site suitability probability distribution, wherein the scatter diagram takes the suitability probability of each established wind power plant as an abscissa and the actually measured installed density corresponding to the established wind power plant as an ordinate.
7. The method as recited in claim 6, further comprising: In the scatter diagram, an X axis is divided into 30 equidistant intervals, points corresponding to 95% of the fractional numbers of the installed density in all sample points are found out in each interval, and a smooth curve is used for connecting all 95% of the fractional number points, wherein the curve is a constraint line.
8. The method as recited in claim 7, further comprising: Calculating the ratio of the actual measured installed density of each established wind power plant to the corresponding theoretical maximum installed density on the constraint line as the residual error of the wind power plant; and for each land utilization type and gradient combination, calculating the average value or the median of the residual errors of all wind power plants under the combination as the utilization coefficient corresponding to the combination.

Description

Wind power installation potential evaluation method based on marginal constraint model Technical Field The invention belongs to the field of maximum developable installed scale evaluation of wind power plants, and particularly relates to a wind power installed potential evaluation method based on a marginal constraint model Background Wind energy has become a core support for constructing a novel energy system by virtue of the advantages of abundant resources, mature technology, large scale development potential and the like. However, the development and utilization of wind energy resources are not easy, and site selection and construction are doubly limited by complex natural geographic conditions and multiple social constraints. Therefore, the regional wind power installation potential is scientifically and accurately evaluated, and the resource endowment, the technical economy and the ecological environment influence are required to be comprehensively prepared, so that a complex system decision process is formed. The main stream evaluation method at present mainly relies on the natural topography and land surface coverage type and evaluates the regional installation potential according to the natural endowment conditions of wind speed, wind power density and other resources. The technical core is that a restriction map is constructed by utilizing weather observation station data, analysis data or numerical simulation results, and the installation potential of different grades is divided by means of space superposition (Overlay) and quantitative analysis functions of a Geographic Information System (GIS) according to preset gradient threshold values and land utilization restriction rules. The method has the advantages of relatively simple and convenient operation and visual result in the aspects of data processing and space visualization. However, the existing evaluation system has significant limitations that firstly, the traditional wind power installation potential evaluation (such as a multi-index weighted superposition model) has the problems of strong subjectivity, unreasonable index weighting, lack of causal interpretation and the like, and is difficult to adapt to higher requirements on scientificity, transparency and adaptability in modern wind power development. Second, the ecological view is lost. The existing method mainly focuses on resources and terrain elements, and generally ignores severe disturbance (such as ecological environment crushing, water and soil loss, noise and electromagnetic interference, landscape cutting and the like) possibly generated by large-scale construction of a wind farm to an regional ecological system. On one hand, the evaluation system is established to be too dependent on subjective assignment and lack of comparison with the existing data, so that the model result is lack of stability and reproducibility, the actual installed capacity and the theoretical installed capacity are too large in difference, and a convincing technical basis is difficult to provide for planning, approval and policy formulation. On the other hand, a large number of initially-judged "suitable" areas are actually overlapped with important ecological protection areas or habitats of rare endangered species to a high degree, projects are forced to be stopped or regulated in the subsequent strict environmental impact Evaluation (EIA) stage due to ecological conflict, so that early investment waste and development process delay are caused, in summary, an intelligent installation potential evaluation method which can integrate multidimensional influence factors and has prediction capability, causality and interpretability is needed to be introduced, and particularly under the large background that ecological protection becomes rigid constraint, the overall coordination and efficient coordination of wind power development and ecological safety are realized. Disclosure of Invention The invention aims to overcome the defects of the prior art and provides a wind power installation potential evaluation method based on a marginal constraint model. The method creatively combines the constraint line model and the random forest algorithm, provides a more accurate and efficient mode for evaluating the installation potential of the wind power plant, and aims to provide decision support for wind energy development. The embodiment of the invention provides a wind power installation potential evaluation method based on a marginal constraint model, which comprises the following steps: constructing a feature vector of a grid corresponding to the wind power plant site selection candidate area; Inputting the feature vector into a preset random forest model to obtain a suitability probability prediction value of wind power development of a grid corresponding to the wind power plant site selection candidate region, and further screening out a final candidate region; Obtaining the theoretical maximum installed densi