A regional wind energy theoretical potential evaluation method, system, terminal and storage medium

Abstract

The application relates to the technical field of data processing, and discloses a regional wind energy theoretical potential evaluation method, a system, a terminal and a storage medium, which comprises the following steps: acquiring multi-source data to generate each grid data set. Weibull parameters are estimated based on historical wind speed, a Gamma prior is constructed as a first prior layer, a Jeffreys super-prior is taken as a second prior layer, and a double-layer Bayesian prior structure is formed. A likelihood function is established based on target year wind speed, and is multiplied by the prior to generate a joint posterior distribution. Positive parameters are converted to an unconstrained space for sampling, and a joint posterior sample set is obtained after quality control. Wind speed correction, air density calculation and effective hour statistics are performed in combination with meteorological and wind turbine parameters, a theoretical wind energy potential posterior sample set is obtained through rotor vertical strip integration, and a posterior mean value, a P90 potential and an uncertainty level are output. The application is suitable for large-scale, high-spatial-resolution regional wind energy resource evaluation and planning management.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method, system, terminal, and computer-readable storage medium for assessing the theoretical potential of regional wind energy. Background Technology

[0002] Regional wind energy resource assessment is a crucial foundation for wind power planning and energy management, and theoretical wind energy potential is a prerequisite for subsequent constraint analyses. Traditional methods typically estimate total resources based on average wind speed or wind power density at a fixed altitude, and identify developable areas by considering land use, slope, and grid connection. In wind speed modeling, the Weibull distribution is widely used due to its simple structure and broad applicability. Common parameter estimation methods include maximum likelihood estimation and moment estimation, which typically estimate distribution parameters as fixed values.

[0003] However, existing theoretical potential assessments have the following shortcomings: they are based solely on wind speed at a single height, neglecting vertical wind speed variations within the turbine rotor sweep range; they use a fixed air density, failing to consider the effects of altitude and temperature variations; they use a fixed number of available hours per year, without statistically analyzing the actual effective operating time for each grid; they do not adequately consider physical constraints such as the Bates limit and turbine power curves; parameter estimation is difficult to quantify uncertainty and cannot integrate historical information from multiple years with observational data from the target year; the assessment results lack a credible range and conservative potential, making it difficult to meet the regional planning requirements for assessing potential stability and risk differences.

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

[0005] The main objective of this invention is to provide a method, system, terminal, and computer-readable storage medium for assessing the theoretical potential of regional wind energy. This invention aims to address the problems of existing technologies that fail to consider variations in vertical wind speed, air density, and effective operating hours of wind turbine rotors, do not adequately consider the physical constraints of wind turbine power extraction, and cannot quantify parameter uncertainties, resulting in insufficient accuracy in assessing the theoretical potential of regional wind energy and a lack of reliable ranges and conservative potential information.

[0006] To achieve the above objectives, the present invention provides a method for assessing the theoretical potential of regional wind energy, the method comprising the following steps: Acquire multi-year historical wind speed data, target year wind speed data, temperature data, air pressure data, digital elevation model data, and wind turbine parameter data for the area to be evaluated. Based on the hub height in the wind turbine parameter data, perform wind speed height correction to adjust the wind speed data to the wind turbine hub height, and generate datasets for each grid. Based on the centralized and corrected multi-year historical wind speed data in the dataset, the Weibull shape parameter and Weibull scale parameter of each grid are estimated by year to obtain the historical shape parameter sequence and historical scale parameter sequence. Based on the historical shape parameter sequence, a Gamma prior distribution of the shape parameter is constructed, and based on the historical scale parameter sequence, a Gamma prior distribution of the scale parameter is constructed to obtain the first prior layer. The Jeffreys prior of the Gamma hyperparameter is used as the second prior layer to construct a two-layer Bayesian prior structure. Based on the centralized and corrected multi-year historical wind speed data in the dataset, a target year wind speed likelihood function is established for each grid. According to Bayes' theorem, the two-layer Bayesian prior structure is multiplied with the target year wind speed likelihood function to generate the joint posterior probability density distribution for each grid. The positive parameters of the joint posterior probability density distribution are transformed to an unconstrained logarithmic space. Based on the transformed probability density and the corresponding gradient, the Hamiltonian Monte Carlo algorithm is used to sample the samples to obtain an initial sample set. Multi-chain convergence test and effective sample size evaluation are performed on the initial sample set to generate a joint posterior sample set of Weibull shape parameter and Weibull scale parameter. Based on the temperature data, air pressure data, digital elevation model data, wind turbine parameter data, and target year wind speed data in the dataset, air density calculation and effective operating hours statistics are performed. Based on the joint posterior sample set, the posterior sample set of the region's theoretical wind energy potential is obtained by rotor vertical strip integration. Based on the posterior sample set, the posterior mean, P90 potential, and uncertainty level of each grid are calculated to obtain the regional wind energy theoretical potential assessment results for the area to be evaluated.

[0007] Optionally, the method for assessing the theoretical potential of regional wind energy, wherein the construction of a Gamma prior distribution of shape parameters based on the historical shape parameter sequence, a Gamma prior distribution of scale parameters based on the historical scale parameter sequence to obtain a first prior layer, and a Jeffreys prior of the Gamma hyperparameter as a second prior layer, to construct a two-layer Bayesian prior structure, specifically includes: The historical shape parameter sequence is used as the shape parameter observation sample. The moment estimation method is used to calculate the sample mean and sample variance of the shape parameter observation sample. Based on the mean and variance formulas of the Gamma distribution, the initial values ​​of the shape hyperparameter and the prior rate hyperparameter are inferred from the sample mean and sample variance of the shape parameter observation sample, and the Gamma prior distribution of the Weibull shape parameter is generated. The historical scale parameter sequence is used as the scale parameter observation sample. The moment estimation method is used to calculate the sample mean and sample variance of the scale parameter observation sample. According to the mean and variance formula of the Gamma distribution, the initial values ​​of the scale hyperparameter and the initial values ​​of the prior rate hyperparameter are inferred from the sample mean and sample variance of the scale parameter observation sample. The Gamma prior distribution of the Weibull scale parameter is generated. The Gamma prior distribution of the Weibull shape parameter and the Gamma prior distribution of the Weibull scale parameter are used as the first prior layer. The shape hyperparameter and prior rate hyperparameter of the Gamma prior distribution of the Weibull shape parameter are used as the first hyperparameter pair, and the shape hyperparameter and prior rate hyperparameter of the Gamma prior distribution of the Weibull scale parameter are used as the second hyperparameter pair. Both the first hyperparameter pair and the second hyperparameter pair are set as joint Jeffreys hyperprior to obtain the second prior layer. Based on the first prior layer and the second prior layer, a two-layer Bayesian prior structure is constructed. In this context, the density function of each joint Jeffreys hyperprior is proportional to the square root of the determinant of the Fisher information matrix of the shape hyperparameter and prior rate hyperparameter of the corresponding Gamma distribution.

[0008] Optionally, the method for assessing the theoretical potential of regional wind energy, wherein the step of multiplying the two-layer Bayesian prior structure with the wind speed likelihood function of the target year according to Bayes' theorem to generate a joint posterior probability density distribution for each grid specifically includes: Obtain the wind speed likelihood function for the target year, the Gamma prior distribution of the Weibull shape parameter, the Gamma prior distribution of the Weibull scale parameter, the joint Jeffreys hyperprior density of the first hyperparameter pair, and the joint Jeffreys hyperprior density of the second hyperparameter pair. Multiply the wind speed likelihood function of the target year, the Gamma prior distribution of the Weibull shape parameter, the Gamma prior distribution of the Weibull scale parameter, and the joint Jeffreys hyperprior density of the two hyperparameter pairs to obtain the joint posterior probability density function expression of the Weibull parameter and the corresponding Gamma hyperparameter of the target year. The joint posterior probability density function expression is used as the joint posterior probability density distribution for each grid.

[0009] Optionally, the method for assessing the theoretical potential of regional wind energy, wherein transforming the positive parameters of the joint posterior probability density distribution to an unconstrained logarithmic space, sampling using the Hamiltonian Monte Carlo algorithm based on the transformed probability density and corresponding gradient to obtain an initial sample set, performing multi-chain convergence testing and effective sample size evaluation on the initial sample set, and generating a joint posterior sample set of Weibull shape parameters and Weibull scale parameters, specifically includes: Set the inter-chain convergence condition and effective sample size threshold, transform the Weibull shape parameter, Weibull scale parameter and the shape hyperparameter and prior rate hyperparameter of the Gamma prior distribution in the joint posterior probability density distribution to the unconstrained logarithmic parameter space to obtain the parameter position vector, and define the potential energy function according to the transformed joint posterior probability density distribution. Among them, the Jacobian correction term for the variable transformation to the unconstrained log-parameter space is automatically included by the posterior sampling procedure when constructing the target probability density of the unconstrained parameter space; An auxiliary momentum variable is introduced into the parameter position vector. A Hamiltonian is constructed through the potential energy function and the auxiliary momentum variable. A dynamic trajectory is generated according to the gradient of the Hamiltonian. Hamiltonian Monte Carlo posterior sampling based on a two-layer Bayesian structure is performed on the dynamic trajectory to obtain an initial sample set. Multiple sampling chains are run in parallel. Based on the Weibull shape parameters and Weibull scale parameters on all sampling chains, as well as the Gamma hyperparameters corresponding to the Weibull shape parameters and Weibull scale parameters, the inter-chain convergence index is calculated. When the inter-chain convergence index of all parameters satisfies the inter-chain convergence condition, the initial sample set of the corresponding spatial grid is determined to have passed the convergence test. The effective sample size is calculated for all parameters that pass the convergence test. When the effective sample size of any parameter to be estimated is lower than the effective sample size threshold, the corresponding grid is determined to have failed the sampling quality test. The posterior sampling is re-executed by increasing the number of preheating times, increasing the number of formal sampling times, or adjusting the sampling control parameters. The samples generated after the preheating period of the sampling chain that passed the quality test are taken as the effective joint posterior sample set.

[0010] Optionally, the method for assessing the theoretical potential of regional wind energy, wherein the step of performing air density calculation and effective operating hours statistics based on the temperature data, air pressure data, digital elevation model data, wind turbine parameter data, and target year wind speed data in the dataset specifically includes: Obtain the hub height, blade radius, cut-in wind speed, and cut-out wind speed from the wind turbine parameter data; The rotor sweep lower edge height and rotor sweep upper edge height are determined based on the hub height and the blade radius, and the vertical interval from the rotor sweep lower edge height to the rotor sweep upper edge height is taken as the rotor sweep vertical interval. The vertical sweeping range of the rotor is divided into multiple micro-element strips, and the center height of each micro-element strip is determined. For each micro-element strip, the air density at the center height of the strip is calculated using the air temperature data, the air pressure data, and the digital elevation model data; Based on the wind speed data for the target year, the cumulative number of hours in each grid where the wind speed falls between the cut-in wind speed and the cut-out wind speed is calculated to obtain the effective operating hours.

[0011] Optionally, the method for assessing the theoretical potential of regional wind energy, wherein obtaining the posterior sample set of the theoretical wind energy potential of the region through rotor vertical strip integration based on the joint posterior sample set specifically includes: Substitute the Weibull shape parameters and Weibull scale parameters of each group in the joint posterior sample set into the Weibull probability density function, and use the wind shear index to correct the wind speed data in the dataset that has been corrected to the hub height to the center height of each micro-element strip. Combine the air density, the effective operating hours, and the cross-sectional area of ​​each micro-element strip, integrate the wind speed interval from the cut-in wind speed to the cut-out wind speed, and accumulate all micro-element strips to generate the theoretical wind energy potential estimate corresponding to the sample. The theoretical wind energy potential estimates corresponding to all samples are used together as the posterior sample set of the regional theoretical wind energy potential.

[0012] Optionally, the method for assessing the theoretical potential of regional wind energy, wherein calculating the posterior mean, P90 potential, and uncertainty level of each grid based on the posterior sample set to obtain the assessment result of the theoretical potential of regional wind energy in the area to be assessed, specifically includes: The 10th percentile of each grid is extracted from the posterior sample set as the P90 potential; Based on the posterior sample set, calculate the posterior mean and 95% confidence interval for each grid, and calculate the difference between the upper and lower limits of the 95% confidence interval. The interval width is obtained based on the difference between the upper and lower limits of the 95% confidence interval. The interval width is then divided into low uncertainty level, medium uncertainty level and high uncertainty level according to a preset classification threshold. The uncertainty level corresponding to each grid is obtained based on the classification result. The posterior mean, the P90 potential, and the uncertainty level corresponding to each grid are used as the regional wind energy theoretical potential assessment results for the area to be evaluated and output.

[0013] Furthermore, to achieve the above objectives, the present invention also provides a regional wind energy theoretical potential assessment system, wherein the regional wind energy theoretical potential assessment system comprises: The multi-source data acquisition module is used to acquire multi-year historical wind speed data, target year wind speed data, temperature data, air pressure data, digital elevation model data, and wind turbine parameter data of the area to be evaluated. Based on the hub height in the wind turbine parameter data, the wind speed height is corrected to the wind turbine hub height, and the datasets of each grid are generated. The wind speed distribution fitting module is used to estimate the Weibull shape parameter and Weibull scale parameter of each grid by year based on the centralized and corrected multi-year historical wind speed data in the dataset, and obtain the historical shape parameter sequence and historical scale parameter sequence. The prior structure construction module is used to construct a Gamma prior distribution of shape parameters based on the historical shape parameter sequence, construct a Gamma prior distribution of scale parameters based on the historical scale parameter sequence, obtain a first prior layer, and use the Jeffreys prior of Gamma hyperparameter as a second prior layer to construct a two-layer Bayesian prior structure. The posterior probability density calculation module is used to establish the target year wind speed likelihood function for each grid based on the centralized and corrected multi-year historical wind speed data of the dataset. According to Bayes' theorem, the two-layer Bayesian prior structure is multiplied with the target year wind speed likelihood function to generate the joint posterior probability density distribution for each grid. The joint posterior sampling module is used to transform the positive parameters of the joint posterior probability density distribution to an unconstrained logarithmic space. Based on the transformed probability density and the corresponding gradient, the Hamilton Monte Carlo algorithm is used to sample to obtain an initial sample set. The initial sample set is then subjected to a multi-chain convergence test and an effective sample size evaluation to generate a joint posterior sample set of Weibull shape parameters and Weibull scale parameters. The wind energy potential calculation module is used to perform air density calculation and effective operating hours statistics based on the temperature data, air pressure data, digital elevation model data, wind turbine parameter data and target year wind speed data in the dataset, and to obtain the posterior sample set of the regional theoretical wind energy potential by rotor vertical strip integration based on the joint posterior sample set. The evaluation and output module is used to calculate the posterior mean, P90 potential, and uncertainty level of each grid based on the posterior sample set, so as to obtain the regional wind energy theoretical potential evaluation result of the area to be evaluated.

[0014] In addition, to achieve the above objectives, the present invention also provides a terminal, wherein the terminal includes: a memory, a processor, and a regional wind energy theoretical potential assessment program stored in the memory and executable on the processor, wherein when the regional wind energy theoretical potential assessment program is executed by the processor, it implements the steps of the regional wind energy theoretical potential assessment method as described above.

[0015] Furthermore, to achieve the above objectives, the present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a regional wind energy theoretical potential assessment program, which, when executed by a processor, implements the steps of the regional wind energy theoretical potential assessment method as described above.

[0016] In this invention, multi-source data of the study area is acquired and preprocessed to generate grid datasets for each grid. Based on multi-year historical wind speed data, the Weibull parameter of each grid is estimated, and a Gamma prior distribution is constructed as the first prior layer, with a Jeffreys super-prior as the second prior layer, forming a two-layer Bayesian prior structure. A likelihood function is established based on the wind speed data of the target year, and multiplied with the prior structure to generate a joint posterior distribution. After transforming the posterior distribution parameters to an unconstrained space, hierarchical Bayesian Monte Carlo sampling is performed, and a joint posterior sample set is obtained through multi-chain convergence testing and effective sample size screening. Combining temperature, air pressure, elevation, and wind turbine parameters, vertical wind speed correction, air density calculation, and effective operating hours statistics are performed. Then, based on the joint posterior sample set, a posterior sample set of theoretical wind energy potential is obtained through rotor vertical strip integration. The posterior mean, P90 potential, and uncertainty level are calculated, and the evaluation results are output. This invention can quantify the uncertainty of Weibull parameters, hyperparameters, and theoretical wind energy potential while retaining historical wind condition information over many years, thereby improving the accuracy of regional wind energy theoretical potential assessment. It is applicable to large-scale, high spatial resolution regional wind energy resource assessment and planning management. Attached Figure Description

[0017] Figure 1 This is a flowchart of a preferred embodiment of the regional wind energy theoretical potential assessment method of the present invention; Figure 2 This is a structural diagram of a preferred embodiment of the regional wind energy theoretical potential assessment system of the present invention; Figure 3 This is a structural diagram of a preferred embodiment of the terminal of the present invention. Detailed Implementation

[0018] This application provides a method, system, and terminal for assessing the theoretical potential of regional wind energy. To make the objectives, technical solutions, and effects of this application clearer and more explicit, the following detailed description is provided with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only for explaining this application and are not intended to limit this application.

[0019] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. It should also be understood that terms such as those defined in general dictionaries should be understood to have the same meaning as in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless specifically defined as herein.

[0020] Furthermore, if the embodiments of this invention involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. If the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed by this invention.

[0021] The regional wind energy theoretical potential assessment method described in the preferred embodiment of the present invention, such as... Figure 1 As shown, the regional wind energy theoretical potential assessment method includes the following steps: Step S10: Obtain multi-year historical wind speed data, target year wind speed data, temperature data, air pressure data, digital elevation model data, and wind turbine parameter data for the area to be evaluated. Correct the wind speed height based on the hub height in the wind turbine parameter data, and generate datasets for each grid.

[0022] In this embodiment, multi-source data of the area to be evaluated is acquired. This multi-source data includes multi-year historical hourly wind speed data, hourly wind speed data for the target year, temperature data, air pressure data, digital elevation model data, and wind turbine parameter data. The wind turbine parameter data includes at least the wind turbine hub height, blade radius, cut-in wind speed, cut-out wind speed, and power coefficient.

[0023] Furthermore, the acquired multi-source data is preprocessed, including coordinate system unification, spatial resolution unification, time scale unification, outlier removal, and missing value processing, dividing the study area into multiple regular spatial grids.

[0024] After preprocessing, the hub height from the wind turbine parameter data is used for wind speed height correction, correcting the wind speed data of each grid from the original observed height to the wind turbine hub height. For the i-th grid, its multi-year historical wind speed data is organized into multiple subsets by year, with each subset corresponding to an hourly wind speed sequence for a historical year. Temperature data, air pressure data, and digital elevation model data are spatially matched with the wind speed data according to grid points, so that each grid point corresponds to a complete set of meteorological sequences and height information.

[0025] Step S20: Based on the centralized and corrected multi-year historical wind speed data in the dataset, estimate the Weibull shape parameter and Weibull scale parameter of each grid by year to obtain the historical shape parameter sequence and historical scale parameter sequence.

[0026] Specifically, based on the multi-year historical hourly wind speed data in the dataset, the Weibull shape parameter and Weibull scale parameter of each spatial grid are estimated by year. For the For each spatial grid, the Weibull probability density function is expressed as: ; The corresponding cumulative distribution function is: ; in, Wind speed is expressed in m / s. This represents the Weibull shape parameter, a dimensionless parameter used to control the concentration and distribution pattern of wind speed. This represents the Weibull scale parameter (m / s), used to reflect the overall scale of the wind speed sequence. It is an exponential function.

[0027] It is understood that the cumulative distribution function is the integral form of the Weibull probability density function. The two describe the random distribution characteristics of wind speed from the perspectives of probability density and cumulative probability, respectively. The Weibull probability density function is used for the subsequent construction of the likelihood function and the integral calculation of the expected value of wind energy, while the cumulative distribution function is used to statistically determine the probability that the wind speed is within a specific interval.

[0028] For the The spatial grid is divided into multiple subsets based on the historical wind speed data over many years. Each subset corresponds to an hourly wind speed sequence for a specific historical year. For each historical year, the shape and scale parameters of the Weibull distribution for that year are estimated. This process is repeated for all historical years to obtain the [missing value]. The estimated sequences of Weibull shape parameters and scale parameters for each spatial grid in all historical years are used as the historical Weibull shape parameter sequence and historical Weibull scale parameter sequence for that grid, respectively. The shape parameter sequence reflects the fluctuation characteristics of the wind speed distribution pattern of the grid in different historical years, while the scale parameter sequence reflects the variation law of the overall wind speed scale of the grid in different historical years.

[0029] Step S30: Construct a Gamma prior distribution of shape parameters based on the historical shape parameter sequence, construct a Gamma prior distribution of scale parameters based on the historical scale parameter sequence to obtain a first prior layer, and use the Jeffreys prior of the Gamma hyperparameter as a second prior layer to construct a two-layer Bayesian prior structure.

[0030] Specifically, the historical shape parameter sequence is used as the shape parameter observation sample. The moment estimation method is used to calculate the sample mean and sample variance of the shape parameter observation sample. Based on the mean and variance formulas of the Gamma distribution, the initial values ​​of the shape hyperparameter and the prior rate hyperparameter are inferred from the sample mean and sample variance of the shape parameter observation sample, and the Gamma prior distribution of the Weibull shape parameter is generated. The historical scale parameter sequence is used as the scale parameter observation sample. The moment estimation method is used to calculate the sample mean and sample variance of the scale parameter observation sample. According to the mean and variance formula of the Gamma distribution, the initial values ​​of the scale hyperparameter and the initial values ​​of the prior rate hyperparameter are inferred from the sample mean and sample variance of the scale parameter observation sample. The Gamma prior distribution of the Weibull scale parameter is generated. The Gamma prior distribution of the Weibull shape parameter and the Gamma prior distribution of the Weibull scale parameter are used as the first prior layer. The shape hyperparameter and prior rate hyperparameter of the Gamma prior distribution of the Weibull shape parameter are used as the first hyperparameter pair, and the shape hyperparameter and prior rate hyperparameter of the Gamma prior distribution of the Weibull scale parameter are used as the second hyperparameter pair. Both the first hyperparameter pair and the second hyperparameter pair are set as joint Jeffreys hyperprior to obtain the second prior layer. Based on the first prior layer and the second prior layer, a two-layer Bayesian prior structure is constructed. In this context, the density function of each joint Jeffreys hyperprior is proportional to the square root of the determinant of the Fisher information matrix of the shape hyperparameter and prior rate hyperparameter of the corresponding Gamma distribution.

[0031] In this embodiment, the first The historical Weibull shape parameter sequence of each grid is used as the shape parameter observation sample. The sample mean and sample variance of this shape parameter observation sample are calculated using the method of moments. The probability density function of the Gamma distribution is: ; in, Represents the positive parameter of the Gamma distribution; Represents the Gamma function; and Let represent the shape parameter and rate parameter of the Gamma distribution, respectively.

[0032] Based on the mean and variance formulas of the Gamma distribution, the initial values ​​of the shape hyperparameters and the prior rate hyperparameters are inversely derived using the sample mean and sample variance of the observed samples of the shape parameters, thus generating the Gamma prior distribution of the Weibull shape parameters: ; in, and These represent the Weibull shape parameters. The shape hyperparameter and prior rate hyperparameter of the Gamma prior distribution.

[0033] Similarly, the first Using the historical Weibull scale parameter sequence of each grid as the scale parameter observation sample, the sample mean and sample variance of this scale parameter observation sample are calculated using the method of moments. Based on the mean and variance formulas of the Gamma distribution, the initial values ​​of the scale hyperparameter and the prior rate hyperparameter are inversely derived using the sample mean and sample variance of the scale parameter observation sample, thus generating the Gamma prior distribution of the Weibull scale parameter. ; in, and These represent the Weibull scale parameters. The shape hyperparameter and prior rate hyperparameter of the Gamma prior distribution.

[0034] The Gamma prior distribution of the generated Weibull shape parameters and the prior distribution of Gamma for the Weibull scale parameter Together they form the first prior layer.

[0035] Furthermore, a second prior layer is constructed. The Gamma prior distribution of the Weibull shape parameters is then... Shape hyperparameters and prior rate hyperparameter and the prior distribution of the Gamma of the Weibull scale parameter. Shape hyperparameters and prior rate hyperparameter All are set to joint Jeffreys super-prior (a type of uninformative prior).

[0036] The density function of the joint Jeffreys hyperprior is: ; in, This represents the joint Jeffreys prior density of the two hyperparameters of the Gamma distribution. The Fisher information matrix represents the parameters of the Gamma distribution. By setting a joint Jeffreys super-prior, the hyperparameters of the Gamma prior no longer depend on fixed empirical values ​​set by humans, reducing the influence of subjective prior settings on the evaluation results. The hyperparameters obtained by adopting the joint Jeffreys super-prior are used as the second prior layer. The first prior layer and the second prior layer together constitute a two-layer Bayesian prior structure.

[0037] Step S40: Based on the centralized and corrected multi-year historical wind speed data in the dataset, establish the target year wind speed likelihood function for each grid. According to Bayes' theorem, multiply the two-layer Bayesian prior structure with the target year wind speed likelihood function to generate the joint posterior probability density distribution for each grid.

[0038] For the i-th spatial grid, its hourly wind speed observations for the target year are arranged in chronological order to form the target year wind speed dataset, where the t-th observation represents the wind speed of the i-th grid at time t in the target year, and n represents the number of valid wind speed observations. The target year wind speed likelihood function is... for: ; Wherein, the preceding equation represents the wind speed likelihood function for the target year of the i-th grid, and the probability density term is the Weibull probability density function, where k and λ represent the Weibull shape parameter and scale parameter for the target year of the grid, respectively.

[0039] The historical Weibull shape parameter sequence and historical Weibull scale parameter sequence obtained in step S20 for the i-th grid are used as historical parameter observation samples, where Y represents the number of historical years. The historical parameter likelihood of the Gamma hyperparameter is established based on the historical parameter sequence. : ; The historical parameter likelihood adopts the Gamma probability density function defined in step S30; the first set of hyperparameters corresponds to the shape hyperparameters and rate hyperparameters of the Weibull shape parameter Gamma prior, respectively, and the second set of hyperparameters corresponds to the shape hyperparameters and rate hyperparameters of the Weibull scale parameter Gamma prior, respectively. The hyperparameter estimates obtained from the historical parameter sequence in step S30 are used for posterior sampling initialization and are not used as fixed hyperparameters.

[0040] Furthermore, according to Bayes' theorem, multiplying the two-layer Bayesian prior structure with the wind speed likelihood function for the target year to generate the joint posterior probability density distribution for each grid cell specifically includes: Obtain the wind speed likelihood function for the target year, the Gamma prior distribution of the Weibull shape parameter, the Gamma prior distribution of the Weibull scale parameter, the joint Jeffreys hyperprior density of the first hyperparameter pair, and the joint Jeffreys hyperprior density of the second hyperparameter pair. Multiply the wind speed likelihood function of the target year, the Gamma prior distribution of the Weibull shape parameter, the Gamma prior distribution of the Weibull scale parameter, and the joint Jeffreys hyperprior density of the two hyperparameter pairs to obtain the joint posterior probability density function expression of the Weibull parameter and the corresponding Gamma hyperparameter of the target year. The joint posterior probability density function expression is used as the joint posterior probability density distribution for each grid.

[0041] In this embodiment, according to Bayes' theorem, the joint posterior probability density distribution of the i-th grid is expressed as: ; The joint posterior distribution includes both the target year's wind speed data and the historical Weibull parameter sequence for the i-th grid. The two Jeffreys hyperprior terms correspond to the shape-rate hyperparameter pairs of the respective Gamma priors for the Weibull shape and scale parameters. Thus, historical year parameter information and target year wind speed observations jointly participate in the posterior update of the current grid.

[0042] Step S50: Transform the positive parameters of the joint posterior probability density distribution to an unconstrained logarithmic space. Based on the transformed probability density and the corresponding gradient, use the Hamiltonian Monte Carlo algorithm to sample and obtain an initial sample set. Perform multi-chain convergence test and effective sample size evaluation on the initial sample set to generate a joint posterior sample set of Weibull shape parameters and Weibull scale parameters.

[0043] Specifically, the inter-chain convergence conditions and effective sample size threshold are set, and the Weibull shape parameter, Weibull scale parameter, and the shape hyperparameter and prior rate hyperparameter of the Gamma prior distribution in the joint posterior probability density distribution are transformed to the unconstrained logarithmic parameter space to obtain the parameter position vector. The potential energy function is defined according to the transformed joint posterior probability density distribution. Among them, the Jacobian correction term for the variable transformation to the unconstrained log-parameter space is automatically included by the posterior sampling procedure when constructing the target probability density of the unconstrained parameter space; An auxiliary momentum variable is introduced into the parameter position vector. A Hamiltonian is constructed through the potential energy function and the auxiliary momentum variable. A dynamic trajectory is generated according to the gradient of the Hamiltonian. Hamiltonian Monte Carlo posterior sampling based on a two-layer Bayesian structure is performed on the dynamic trajectory to obtain an initial sample set. Multiple sampling chains are run in parallel. Based on the Weibull shape parameters and Weibull scale parameters on all sampling chains, as well as the Gamma hyperparameters corresponding to the Weibull shape parameters and Weibull scale parameters, the inter-chain convergence index is calculated. When the inter-chain convergence index of all parameters satisfies the inter-chain convergence condition, the initial sample set of the corresponding spatial grid is determined to have passed the convergence test. The effective sample size is calculated for all parameters that pass the convergence test. When the effective sample size of any parameter to be estimated is lower than the effective sample size threshold, the corresponding grid is determined to have failed the sampling quality test. The posterior sampling is re-executed by increasing the number of preheating times, increasing the number of formal sampling times, or adjusting the sampling control parameters. The samples generated after the preheating period of the sampling chain that passed the quality test are taken as the effective joint posterior sample set.

[0044] It is understandable that the joint posterior probability density distribution typically lacks an analytical solution, making it impossible to directly obtain the posterior distribution of the parameters analytically. Therefore, the Hamiltonian Monte Carlo (HMC) method is employed for posterior sampling. All positive parameters in the joint posterior probability density distribution are transformed to an unconstrained logarithmic parameter space. Since the Weibull shape parameter, Weibull scale parameter, and the shape and prior rate hyperparameters of the Gamma prior distribution are all positive parameters, direct sampling in a constrained space would be limited by boundaries, affecting sampling efficiency. Therefore, a logarithmic transformation is performed on all positive parameters, mapping them to an unconstrained real space. The transformed parameters include the Weibull shape parameter, Weibull scale parameter, shape and prior rate hyperparameters of the shape parameter Gamma prior distribution, and shape and prior rate hyperparameters of the scale parameter Gamma prior distribution for the i-th grid. This invention uses the pymc (a probability programming library) program. The transformed posterior density needs to take into account the Jacobian term of the variable transformation. The Jacobian correction term of the variable transformation to the unconstrained logarithmic parameter space is automatically included by the posterior sampling program when constructing the target probability density of the unconstrained parameter space.

[0045] Furthermore, a potential energy function is defined based on the transformed joint posterior probability density distribution. This function takes a negative logarithmic form, ensuring that regions with higher posterior probability densities have lower potential energy, corresponding to the tendency of particles in physical systems to prefer low-potential states. An auxiliary momentum variable is introduced for the parameter position vector. This auxiliary momentum variable follows a normal distribution with zero mean, and its covariance matrix is ​​determined by the positive definite mass matrix. The introduction of the auxiliary momentum variable transforms the parameter inference problem into a simulation problem of a Hamiltonian dynamics system, enabling the sampling process to utilize gradient information to generate high-acceptance-rate proposal samples far from the current state. A Hamiltonian is constructed using the potential energy function and the auxiliary momentum variable. The Hamiltonian consists of the potential energy term and the kinetic energy term, constituting the total energy of the system. The kinetic energy term is calculated from the auxiliary momentum variable and the mass matrix. A dynamic trajectory is generated based on the gradient of the Hamiltonian, and a parameter trajectory is constructed using the target posterior probability density gradient. By simulating the particle's motion trajectory in the potential energy field, efficient exploration of the parameter space can be achieved while maintaining a high acceptance rate. Based on the dynamic trajectory, Hamiltonian Monte Carlo posterior sampling based on a two-layer Bayesian structure is performed, and an initial sample set is obtained after multiple iterations.

[0046] Furthermore, to ensure the reliability and statistical validity of the sampling results, a multi-chain convergence test and effective sample size evaluation are performed on the initial sample set. Inter-chain convergence conditions and an effective sample size threshold are set, and at least four sampling chains are run in parallel. Inter-chain convergence indices for the Weibull shape parameter and Weibull scale parameter on each sampling chain are calculated. The inter-chain convergence index assesses whether multiple sampling chains converge to the same target distribution by comparing the inter-chain variance with the intra-chain variance. When the inter-chain convergence index is less than the inter-chain convergence condition, the initial sample set of the corresponding spatial grid is determined to have passed the inter-chain convergence test, indicating that the sampling chain has reached a steady-state distribution.

[0047] The effective sample size is calculated for the parameters to be estimated in the sample set that have passed the inter-chain convergence test. The effective sample size is the statistic corresponding to the entire set of related samples of a certain parameter. When the effective sample size of any parameter to be estimated is lower than the effective sample size threshold, the corresponding grid is determined to have failed the sampling quality test. The posterior sampling is re-executed by increasing the number of warm-up times, increasing the number of formal sampling times, or adjusting the sampling control parameters. The samples generated after the warm-up period of the sampling chain that has passed the quality test are taken as the effective joint posterior sample set.

[0048] Specifically, for the first A spatial grid, with parameters to be estimated including Weibull shape parameters. Weibull scale parameters The parameters to be estimated, including the shape hyperparameter and rate hyperparameter, control the Gamma prior distributions of both parameters. Since all the parameters to be estimated are positive, when using the HMC sampling program for calculation, an internal transformation suitable for the unconstrained parameter space is performed on these positive parameters, and the transformed target probability density is corrected accordingly. The parameter transformation and the corresponding probability density correction can be automatically completed by the probabilistic programming program when constructing the HMC target posterior distribution.

[0049] The potential function is defined based on the transformed joint posterior probability density. : ; in, This indicates that the HMC sampling process corresponds to the first... The parameter vector of each parameter to be estimated in each grid cell; Indicates the first Each grid is used for posterior inference of wind speed data and historical parameter information; Indicates the first Joint posterior probability density of the target in each grid; This represents the potential function determined by the joint posterior probability density of the target.

[0050] Introduce auxiliary momentum variables to the parameter vector : ; And construct the Hamiltonian : ; in, Represents the auxiliary momentum vector; Represents the positive definite mass matrix; It represents the Hamiltonian, which consists of potential energy and kinetic energy terms.

[0051] HMC constructs Hamiltonian dynamic trajectories using the gradient of the joint posterior probability density of the target, generates candidate states in the parameter space and performs posterior sampling, thereby obtaining joint posterior samples of the Weibull shape parameters, scale parameters and related Gamma hyperparameters of the i-th grid.

[0052] For each spatial grid, multiple independent sampling chains are run in parallel. For the Weibull shape parameter, Weibull scale parameter, and corresponding Gamma hyperparameter, the inter-chain convergence index and effective sample size are calculated based on all sampling chains. When all parameters to be estimated meet the preset inter-chain convergence and effective sample size conditions, the grid is deemed to have passed the posterior sampling quality test. For grids that fail the quality test, posterior sampling can be re-executed by increasing the number of preheating iterations, increasing the number of formal sampling iterations, or adjusting the HMC sampling control parameters.

[0053] The samples that pass the sampling quality test after the preheating phase are used as the effective joint posterior sample set for the grid. For the first... The joint posterior sample representation of the Weibull shape parameter and scale parameter of each grid is as follows: ; in, Indicates the first The number of valid posterior samples retained after each grid passes quality inspection; This indicates the posterior sample number in the joint posterior distribution of the grid; and They represent the first The first grid The Weibull shape and scale parameters corresponding to the set of posterior samples. These effective joint posterior samples are used for subsequent uncertainty calculations of wind speed distribution and theoretical wind energy potential.

[0054] Step S60: Based on the temperature data, air pressure data, digital elevation model data, wind turbine parameter data, and target year wind speed data in the dataset, perform air density calculation and effective operating hours statistics, and obtain the posterior sample set of the regional theoretical wind energy potential by rotor vertical strip integration based on the joint posterior sample set.

[0055] The process of performing wind speed and height correction, air density calculation, and effective operating hours statistics based on the temperature data, air pressure data, digital elevation model data, wind turbine parameter data, and target year wind speed data in the dataset specifically includes: Obtain the hub height, blade radius, cut-in wind speed, and cut-out wind speed from the wind turbine parameter data; The rotor sweep lower edge height and rotor sweep upper edge height are determined based on the hub height and the blade radius, and the vertical interval from the rotor sweep lower edge height to the rotor sweep upper edge height is taken as the rotor sweep vertical interval. The vertical sweeping range of the rotor is divided into multiple micro-element strips, and the center height of each micro-element strip is determined. For each micro-element strip, the air density at the center height of the strip is calculated using the air temperature data, the air pressure data, and the digital elevation model data; Based on the wind speed data for the target year, the cumulative number of hours in each grid where the wind speed falls between the cut-in wind speed and the cut-out wind speed is calculated to obtain the effective operating hours.

[0056] In this embodiment, each effective spatial grid is treated as an independent evaluation location, and the theoretical maximum wind energy potential of that location is calculated using a virtual representative wind turbine of the same specification. This calculation does not consider the turbine spacing, wake loss, or actual turbine placement, and is only used to compare the theoretical wind energy potential of each spatial grid under the same representative wind turbine technology conditions.

[0057] Let the height of the representative wind turbine hub be H and the blade radius be R, then the vertical sweep interval of the rotor is ( ),in, ; Divide the vertical interval into J infinitesimal strips, each strip having a corresponding center height denoted as Δz. Then the area of ​​the j-th infinitesimal strip is approximately... : ; in, This represents the center height of the j-th micro-element strip. Based on the wind speed already corrected to the hub height H in step S10, the wind speed is further corrected to the center height of the j-th strip using the wind shear index corresponding to the i-th grid. : ; in, This represents the wind shear index corresponding to the i-th grid. When using the above proportional height correction, the Weibull shape parameter remains unchanged, while the scale parameter is adjusted by the same proportion. For the m-th posterior parameter, we have: ; Where m represents the group index of the posterior sample. Let represent the Weibull scale parameter of the i-th grid and the m-th posterior sample at the hub height. Let represent the Weibull shape parameters of the i-th grid and the m-th posterior sample at the hub height. This represents the Weibull shape parameter after correction to the center height of the j-th strip. This represents the Weibull scale parameter after correction to the center height of the j-th strip.

[0058] Air density at the center height of the j-th strip in the i-th grid Calculate using the following formula: ; in, This represents the air pressure at the center height of the i-th grid and the j-th strip. This represents the gas constant of dry air. This represents the absolute temperature at the center height of the i-th grid and j-th strip. The air pressure in the numerator and the absolute temperature in the denominator are taken as the corresponding values ​​at the center height of the j-th strip in the i-th grid, respectively. The dry air gas constant is taken as 287 J / (kg·K). Based on the hourly wind speed corrected to hub height in step S10 for the target year, the effective operating hours of the i-th grid are calculated. : ; in, and These represent the cut-out wind speed and the cut-in wind speed, respectively. ( ) is an indicator function, which takes the value 1 if the condition is true and 0 otherwise; for hourly data, Δt takes the value of 1 hour.

[0059] Furthermore, the step of obtaining the posterior sample set of the regional theoretical wind energy potential through rotor vertical strip integration based on the joint posterior sample set specifically includes: Substitute the Weibull shape parameters and Weibull scale parameters of each group in the joint posterior sample set into the Weibull probability density function, and use the wind shear index to correct the wind speed data in the dataset that has been corrected to the hub height to the center height of each micro-element strip. Combine the air density, the effective operating hours, and the cross-sectional area of ​​each micro-element strip, integrate the wind speed interval from the cut-in wind speed to the cut-out wind speed, and accumulate all micro-element strips to generate the theoretical wind energy potential estimate corresponding to the sample. The theoretical wind energy potential estimates corresponding to all samples are used together as the posterior sample set of the regional theoretical wind energy potential.

[0060] Understandably, since the effective operating hours have already been counted separately, to avoid repeatedly including the probability of wind speed falling into the cut-in / cut-out interval through probability integration, the conditional expectation of the cube of the wind speed within that operating interval is used for each set of posterior parameters m. : ; in, This is the Weibull cumulative distribution function. When the denominator is 0, the wind energy potential of the corresponding strip is not included in the grid.

[0061] This invention is used to calculate the theoretical maximum wind energy potential, with the power factor of the representative wind turbine adopting the Bates limit of 16 / 27. For the m-th set of posterior parameters in the i-th grid, the estimated theoretical wind energy potential of that grid is: ; in, This represents the theoretical wind energy potential estimate for the i-th grid under the m-th set of posterior parameters. This represents the total number of strips. By traversing all valid posterior samples of the grid, the posterior sample set of the theoretical wind energy potential of the grid can be formed; repeat the above calculation for all spatial grids to obtain the posterior spatial distribution of the theoretical wind energy potential of the area to be evaluated.

[0062] Step S70: Calculate the posterior mean, P90 potential, and uncertainty level of each grid based on the posterior sample set to obtain the regional wind energy theoretical potential assessment result of the area to be evaluated.

[0063] Specifically, the 10th percentile of each grid is extracted from the posterior sample set as the P90 potential; Based on the posterior sample set, calculate the posterior mean and 95% confidence interval for each grid, and calculate the difference between the upper and lower limits of the 95% confidence interval. The interval width is obtained based on the difference between the upper and lower limits of the 95% confidence interval. The interval width is then divided into low uncertainty level, medium uncertainty level and high uncertainty level according to a preset classification threshold. The uncertainty level corresponding to each grid is obtained based on the classification result. The posterior mean, the P90 potential, and the uncertainty level corresponding to each grid are used as the regional wind energy theoretical potential assessment results for the area to be evaluated and output.

[0064] Understandably, the posterior mean is the arithmetic mean of all posterior samples, reflecting the expected level of the theoretical wind energy potential of the grid; the 95% confidence interval is composed of the 2.5% quantile and the 97.5% quantile of the posterior samples, indicating that the theoretical wind energy potential has a 95% probability of falling within this interval, and is used to quantify the range of uncertainty in the assessment.

[0065] In this embodiment, the 10th percentile of each grid is extracted from the posterior sample set as the P90 potential, which means that there is a 90% probability that the theoretical wind energy potential exceeds this value. This represents the wind energy potential level under a conservative scenario and is suitable for risk-averse planning decision-making scenarios. The interval width is calculated based on the difference between the upper and lower limits of the 95% confidence interval. The interval width reflects the degree of uncertainty of the grid assessment result. The larger the width, the greater the fluctuation range of the wind energy potential estimate in the area, and the lower the reliability of the assessment result; the smaller the width, the more stable the estimate. The interval width is divided into low uncertainty, medium uncertainty, and high uncertainty levels according to a preset classification threshold. For example, grids with an interval width less than the first threshold are marked as low uncertainty level, grids with an interval width greater than the second threshold are marked as high uncertainty level, and those in between are marked as medium uncertainty level. The preset classification threshold can be adaptively set according to the overall uncertainty distribution characteristics of the study area, or it can be pre-specified by the assessment personnel according to management needs. In another embodiment, wind energy potential results at different confidence levels such as P50 and P75 can also be output to provide conservative, average, and optimistic scenarios to support regional wind energy development.

[0066] The posterior mean, the P90 potential, and the determined uncertainty level are used as the output of the regional wind energy theoretical potential assessment result for the area to be assessed. Combining these three factors expands the assessment result from a single deterministic value to a comprehensive assessment result that includes expected levels, conservative scenarios, and reliability information, providing support for regional wind energy resource management, determining wind farm site selection priorities, and analyzing development scenarios under different confidence levels.

[0067] Specifically, for the i-th grid, its theoretical wind energy potential posterior mean is... for: ; The P90 potential is taken as the 10th quantile of the posterior distribution of the theoretical wind energy potential of this grid: ; The 95% posterior confidence interval is formed by the 2.5% quantile and the 97.5% quantile of the posterior distribution of this grid: ; Among them, the quantile function This represents the value of the posterior distribution of the theoretical wind energy potential of the corresponding grid at a given probability location. The width of the 95% posterior confidence interval is used to characterize the uncertainty of the grid. ; in, The degree of uncertainty is used to characterize the posterior uncertainty of the i-th grid. Each grid is divided into low, medium, and high uncertainty levels according to a preset grading threshold, or determined based on the overall distribution of the widths of all grid intervals in the study area.

[0068] The posterior mean, P90 potential, and uncertainty level of each grid are output as the regional wind energy theoretical potential assessment results. P90 indicates that there is approximately a 90% posterior probability that the theoretical wind energy potential is not lower than this value, which belongs to the exceedance probability index; in other embodiments, theoretical wind energy potential results at different exceedance probability levels such as P50 and P75 can also be output.

[0069] Furthermore, such as Figure 2 As shown, based on the above-mentioned method for assessing the theoretical potential of regional wind energy, the present invention also provides a system for assessing the theoretical potential of regional wind energy, wherein the system comprises: The multi-source data acquisition module 51 is used to acquire multi-year historical wind speed data, target year wind speed data, temperature data, air pressure data, digital elevation model data and wind turbine parameter data of the area to be evaluated. Based on the hub height in the wind turbine parameter data, the wind speed height is corrected to the wind turbine hub height, and the dataset of each grid is generated. The wind speed distribution fitting module 52 is used to estimate the Weibull shape parameter and Weibull scale parameter of each grid by year based on the centralized and corrected multi-year historical wind speed data in the dataset, and obtain the historical shape parameter sequence and historical scale parameter sequence. The prior structure construction module 53 is used to construct the Gamma prior distribution of shape parameters according to the historical shape parameter sequence, construct the Gamma prior distribution of scale parameters according to the historical scale parameter sequence, obtain the first prior layer, and use the Jeffreys prior of Gamma hyperparameter as the second prior layer to construct a two-layer Bayesian prior structure. The posterior probability density calculation module 54 is used to establish the target year wind speed likelihood function for each grid based on the centralized and corrected multi-year historical wind speed data of the dataset. According to Bayes' theorem, the two-layer Bayesian prior structure is multiplied with the target year wind speed likelihood function to generate the joint posterior probability density distribution for each grid. The joint posterior sampling module 55 is used to transform the positive parameters of the joint posterior probability density distribution to an unconstrained logarithmic space, and to sample using the Hamilton Monte Carlo algorithm based on the transformed probability density and the corresponding gradient to obtain an initial sample set. The initial sample set is then subjected to a multi-chain convergence test and an effective sample size evaluation to generate a joint posterior sample set of Weibull shape parameters and Weibull scale parameters. The wind energy potential calculation module 56 is used to perform air density calculation and effective operating hours statistics based on the temperature data, air pressure data, digital elevation model data, wind turbine parameter data and target year wind speed data in the dataset, and to obtain the posterior sample set of the regional theoretical wind energy potential by rotor vertical strip integration based on the joint posterior sample set. The evaluation and output module 57 is used to calculate the posterior mean, P90 potential and uncertainty level of each grid based on the posterior sample set, so as to obtain the regional wind energy theoretical potential evaluation result of the area to be evaluated.

[0070] Furthermore, such as Figure 3 As shown, based on the above-mentioned regional wind energy theoretical potential assessment method and system, the present invention also provides a terminal, which includes a processor 10, a memory 20 and a display 30. Figure 3 Only some of the terminal components are shown; however, it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0071] In some embodiments, the memory 20 may be an internal storage unit of the terminal, such as a hard disk or memory. In other embodiments, the memory 20 may be an external storage device of the terminal, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, etc., equipped on the terminal. Further, the memory 20 may include both internal and external storage devices. The memory 20 is used to store application software and various types of data installed on the terminal, such as the program code installed on the terminal. The memory 20 can also be used to temporarily store data that has been output or will be output. In one embodiment, the memory 20 stores a regional wind energy theoretical potential assessment program 40, which can be executed by the processor 10 to implement the regional wind energy theoretical potential assessment method of this application.

[0072] In some embodiments, the processor 10 may be a central processing unit (CPU), a microprocessor, or other data processing chip, used to run program code stored in the memory 20 or process data, such as executing the regional wind energy theoretical potential assessment method.

[0073] In some embodiments, the display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. The display 30 is used to display information on the terminal and to display a visual user interface. The components of the terminal communicate with each other via a system bus.

[0074] In one embodiment, when the processor 10 executes the regional wind energy theoretical potential assessment program 40 in the memory 20, the following steps are performed: Acquire multi-year historical wind speed data, target year wind speed data, temperature data, air pressure data, digital elevation model data, and wind turbine parameter data for the area to be evaluated. Based on the hub height in the wind turbine parameter data, perform wind speed height correction to adjust the wind speed data to the wind turbine hub height, and generate datasets for each grid. Based on the centralized and corrected multi-year historical wind speed data in the dataset, the Weibull shape parameter and Weibull scale parameter of each grid are estimated by year to obtain the historical shape parameter sequence and historical scale parameter sequence. Based on the historical shape parameter sequence, a Gamma prior distribution of the shape parameter is constructed, and based on the historical scale parameter sequence, a Gamma prior distribution of the scale parameter is constructed to obtain the first prior layer. The Jeffreys prior of the Gamma hyperparameter is used as the second prior layer to construct a two-layer Bayesian prior structure. Based on the centralized and corrected multi-year historical wind speed data in the dataset, a target year wind speed likelihood function is established for each grid. According to Bayes' theorem, the two-layer Bayesian prior structure is multiplied with the target year wind speed likelihood function to generate the joint posterior probability density distribution for each grid. The positive parameters of the joint posterior probability density distribution are transformed to an unconstrained logarithmic space. Based on the transformed probability density and the corresponding gradient, the Hamiltonian Monte Carlo algorithm is used to sample the samples to obtain an initial sample set. Multi-chain convergence test and effective sample size evaluation are performed on the initial sample set to generate a joint posterior sample set of Weibull shape parameter and Weibull scale parameter. Based on the temperature data, air pressure data, digital elevation model data, wind turbine parameter data, and target year wind speed data in the dataset, air density calculation and effective operating hours statistics are performed. Based on the joint posterior sample set, the posterior sample set of the region's theoretical wind energy potential is obtained by rotor vertical strip integration. Based on the posterior sample set, the posterior mean, P90 potential, and uncertainty level of each grid are calculated to obtain the regional wind energy theoretical potential assessment results for the area to be evaluated.

[0075] Specifically, the construction of a Gamma prior distribution for shape parameters based on the historical shape parameter sequence, and a Gamma prior distribution for scale parameters based on the historical scale parameter sequence, to obtain a first prior layer, and using the Jeffreys prior of the Gamma hyperparameter as a second prior layer, constitutes a two-layer Bayesian prior structure, which includes: The historical shape parameter sequence is used as the shape parameter observation sample. The moment estimation method is used to calculate the sample mean and sample variance of the shape parameter observation sample. Based on the mean and variance formulas of the Gamma distribution, the initial values ​​of the shape hyperparameter and the prior rate hyperparameter are inferred from the sample mean and sample variance of the shape parameter observation sample, and the Gamma prior distribution of the Weibull shape parameter is generated. The historical scale parameter sequence is used as the scale parameter observation sample. The moment estimation method is used to calculate the sample mean and sample variance of the scale parameter observation sample. According to the mean and variance formula of the Gamma distribution, the initial values ​​of the scale hyperparameter and the initial values ​​of the prior rate hyperparameter are inferred from the sample mean and sample variance of the scale parameter observation sample. The Gamma prior distribution of the Weibull scale parameter is generated. The Gamma prior distribution of the Weibull shape parameter and the Gamma prior distribution of the Weibull scale parameter are used as the first prior layer. The shape hyperparameter and prior rate hyperparameter of the Gamma prior distribution of the Weibull shape parameter are used as the first hyperparameter pair, and the shape hyperparameter and prior rate hyperparameter of the Gamma prior distribution of the Weibull scale parameter are used as the second hyperparameter pair. Both the first hyperparameter pair and the second hyperparameter pair are set as joint Jeffreys hyperprior to obtain the second prior layer. Based on the first prior layer and the second prior layer, a two-layer Bayesian prior structure is constructed. In this context, the density function of each joint Jeffreys hyperprior is proportional to the square root of the determinant of the Fisher information matrix of the shape hyperparameter and prior rate hyperparameter of the corresponding Gamma distribution.

[0076] Specifically, the step of multiplying the two-layer Bayesian prior structure with the wind speed likelihood function of the target year according to Bayes' theorem to generate the joint posterior probability density distribution for each grid includes: Obtain the wind speed likelihood function for the target year, the Gamma prior distribution of the Weibull shape parameter, the Gamma prior distribution of the Weibull scale parameter, the joint Jeffreys hyperprior density of the first hyperparameter pair, and the joint Jeffreys hyperprior density of the second hyperparameter pair. Multiply the wind speed likelihood function of the target year, the Gamma prior distribution of the Weibull shape parameter, the Gamma prior distribution of the Weibull scale parameter, and the joint Jeffreys hyperprior density of the two hyperparameter pairs to obtain the joint posterior probability density function expression of the Weibull parameter and the corresponding Gamma hyperparameter of the target year. The joint posterior probability density function expression is used as the joint posterior probability density distribution for each grid.

[0077] Specifically, the process of transforming the positive parameters of the joint posterior probability density distribution to an unconstrained logarithmic space, sampling using the Hamiltonian Monte Carlo algorithm based on the transformed probability density and corresponding gradient to obtain an initial sample set, performing multi-chain convergence testing and effective sample size evaluation on the initial sample set, and generating a joint posterior sample set for the Weibull shape parameter and the Weibull scale parameter includes: Set the inter-chain convergence condition and effective sample size threshold, transform the Weibull shape parameter, Weibull scale parameter and the shape hyperparameter and prior rate hyperparameter of the Gamma prior distribution in the joint posterior probability density distribution to the unconstrained logarithmic parameter space to obtain the parameter position vector, and define the potential energy function according to the transformed joint posterior probability density distribution. Among them, the Jacobian correction term for the variable transformation to the unconstrained log-parameter space is automatically included by the posterior sampling procedure when constructing the target probability density of the unconstrained parameter space; An auxiliary momentum variable is introduced into the parameter position vector. A Hamiltonian is constructed through the potential energy function and the auxiliary momentum variable. A dynamic trajectory is generated according to the gradient of the Hamiltonian. Hamiltonian Monte Carlo posterior sampling based on a two-layer Bayesian structure is performed on the dynamic trajectory to obtain an initial sample set. Multiple sampling chains are run in parallel. Based on the Weibull shape parameters and Weibull scale parameters on all sampling chains, as well as the Gamma hyperparameters corresponding to the Weibull shape parameters and Weibull scale parameters, the inter-chain convergence index is calculated. When the inter-chain convergence index of all parameters satisfies the inter-chain convergence condition, the initial sample set of the corresponding spatial grid is determined to have passed the convergence test. The effective sample size is calculated for all parameters that pass the convergence test. When the effective sample size of any parameter to be estimated is lower than the effective sample size threshold, the corresponding grid is determined to have failed the sampling quality test. The posterior sampling is re-executed by increasing the number of preheating times, increasing the number of formal sampling times, or adjusting the sampling control parameters. The samples generated after the preheating period of the sampling chain that passed the quality test are taken as the effective joint posterior sample set.

[0078] Specifically, the step of performing air density calculation and effective operating hours statistics based on the temperature data, air pressure data, digital elevation model data, wind turbine parameter data, and target year wind speed data in the dataset includes: Obtain the hub height, blade radius, cut-in wind speed, and cut-out wind speed from the wind turbine parameter data; The rotor sweep lower edge height and rotor sweep upper edge height are determined based on the hub height and the blade radius, and the vertical interval from the rotor sweep lower edge height to the rotor sweep upper edge height is taken as the rotor sweep vertical interval. The vertical sweeping range of the rotor is divided into multiple micro-element strips, and the center height of each micro-element strip is determined. For each micro-element strip, the air density at the center height of the strip is calculated using the air temperature data, the air pressure data, and the digital elevation model data; Based on the wind speed data for the target year, the cumulative number of hours in each grid where the wind speed falls between the cut-in wind speed and the cut-out wind speed is calculated to obtain the effective operating hours.

[0079] Specifically, obtaining the posterior sample set of the regional theoretical wind energy potential through rotor vertical strip integration based on the joint posterior sample set includes: Substitute the Weibull shape parameters and Weibull scale parameters of each group in the joint posterior sample set into the Weibull probability density function, and use the wind shear index to correct the wind speed data in the dataset that has been corrected to the hub height to the center height of each micro-element strip. Combine the air density, the effective operating hours, and the cross-sectional area of ​​each micro-element strip, integrate the wind speed interval from the cut-in wind speed to the cut-out wind speed, and accumulate all micro-element strips to generate the theoretical wind energy potential estimate corresponding to the sample. The theoretical wind energy potential estimates corresponding to all samples are used together as the posterior sample set of the regional theoretical wind energy potential.

[0080] Specifically, the step of calculating the posterior mean, P90 potential, and uncertainty level of each grid based on the posterior sample set to obtain the regional wind energy theoretical potential assessment result for the area to be assessed includes: The 10th percentile of each grid is extracted from the posterior sample set as the P90 potential; Based on the posterior sample set, calculate the posterior mean and 95% confidence interval for each grid, and calculate the difference between the upper and lower limits of the 95% confidence interval. The interval width is obtained based on the difference between the upper and lower limits of the 95% confidence interval. The interval width is then divided into low uncertainty level, medium uncertainty level and high uncertainty level according to a preset classification threshold. The uncertainty level corresponding to each grid is obtained based on the classification result. The posterior mean, the P90 potential, and the uncertainty level corresponding to each grid are used as the regional wind energy theoretical potential assessment results for the area to be evaluated and output.

[0081] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a regional wind energy theoretical potential assessment program, which, when executed by a processor, implements the steps of the regional wind energy theoretical potential assessment method as described above.

[0082] In summary, this invention provides a method, system, terminal, and storage medium for assessing the theoretical potential of regional wind energy. The method includes: acquiring multi-source data of the study area and generating grid datasets after preprocessing; estimating the Weibull parameters of each grid based on multi-year historical wind speed data, constructing a Gamma prior distribution as the first prior layer, and using a Jeffreys super-prior as the second prior layer to form a two-layer Bayesian prior structure; establishing a likelihood function based on wind speed data of the target year, multiplying it with the prior structure to generate a joint posterior distribution; transforming the posterior distribution parameters to an unconstrained space and performing hierarchical Bayesian Monte Carlo sampling, obtaining a joint posterior sample set through multi-chain convergence testing and effective sample size screening; combining air temperature, air pressure, elevation, and wind turbine parameters, performing vertical wind speed correction, air density calculation, and effective operating hours statistics, and then obtaining the posterior sample set of theoretical wind energy potential through rotor vertical strip integration based on the joint posterior sample set; calculating the posterior mean, P90 potential, and uncertainty level, and outputting the assessment results. This invention quantifies uncertainty while preserving historical wind condition information. This invention can quantify the uncertainty of Weibull parameters, hyperparameters, and theoretical wind energy potential while retaining historical wind condition information over many years, thereby improving the accuracy of regional wind energy theoretical potential assessment. It is applicable to large-scale, high spatial resolution regional wind energy resource assessment and planning management.

[0083] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal that includes that element.

[0084] Of course, those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0085] It should be understood that the application of the present invention is not limited to the examples above. Those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. A method for assessing the theoretical potential of regional wind energy, characterized in that, The aforementioned regional wind energy theoretical potential assessment method includes: Acquire multi-year historical wind speed data, target year wind speed data, temperature data, air pressure data, digital elevation model data, and wind turbine parameter data for the area to be evaluated. Based on the hub height in the wind turbine parameter data, perform wind speed height correction to adjust the wind speed data to the wind turbine hub height, and generate datasets for each grid. Based on the centralized and corrected multi-year historical wind speed data in the dataset, the Weibull shape parameter and Weibull scale parameter of each grid are estimated by year to obtain the historical shape parameter sequence and historical scale parameter sequence. Based on the historical shape parameter sequence, a Gamma prior distribution of the shape parameter is constructed, and based on the historical scale parameter sequence, a Gamma prior distribution of the scale parameter is constructed to obtain the first prior layer. The Jeffreys prior of the Gamma hyperparameter is used as the second prior layer to construct a two-layer Bayesian prior structure. Based on the centralized and corrected multi-year historical wind speed data in the dataset, a target year wind speed likelihood function is established for each grid. According to Bayes' theorem, the two-layer Bayesian prior structure is multiplied with the target year wind speed likelihood function to generate the joint posterior probability density distribution for each grid. The positive parameters of the joint posterior probability density distribution are transformed to an unconstrained logarithmic space. Based on the transformed probability density and the corresponding gradient, the Hamiltonian Monte Carlo algorithm is used to sample the samples to obtain an initial sample set. Multi-chain convergence test and effective sample size evaluation are performed on the initial sample set to generate a joint posterior sample set of Weibull shape parameter and Weibull scale parameter. Based on the temperature data, air pressure data, digital elevation model data, wind turbine parameter data, and target year wind speed data in the dataset, air density calculation and effective operating hours statistics are performed. Based on the joint posterior sample set, the posterior sample set of the region's theoretical wind energy potential is obtained by rotor vertical strip integration. Based on the posterior sample set, the posterior mean, P90 potential, and uncertainty level of each grid are calculated to obtain the regional wind energy theoretical potential assessment results for the area to be evaluated.

2. The method for assessing the theoretical potential of regional wind energy according to claim 1, characterized in that, The construction of a Gamma prior distribution for shape parameters based on the historical shape parameter sequence, and a Gamma prior distribution for scale parameters based on the historical scale parameter sequence, yields a first prior layer. A second prior layer is constructed using the Jeffreys prior of the Gamma hyperparameter as the second prior layer, thus forming a two-layer Bayesian prior structure. Specifically, this includes: The historical shape parameter sequence is used as the shape parameter observation sample. The moment estimation method is used to calculate the sample mean and sample variance of the shape parameter observation sample. Based on the mean and variance formulas of the Gamma distribution, the initial values ​​of the shape hyperparameter and the prior rate hyperparameter are inferred from the sample mean and sample variance of the shape parameter observation sample, and the Gamma prior distribution of the Weibull shape parameter is generated. The historical scale parameter sequence is used as the scale parameter observation sample. The moment estimation method is used to calculate the sample mean and sample variance of the scale parameter observation sample. According to the mean and variance formula of the Gamma distribution, the initial values ​​of the scale hyperparameter and the initial values ​​of the prior rate hyperparameter are inferred from the sample mean and sample variance of the scale parameter observation sample. The Gamma prior distribution of the Weibull scale parameter is generated. The Gamma prior distribution of the Weibull shape parameter and the Gamma prior distribution of the Weibull scale parameter are used as the first prior layer. The shape hyperparameter and prior rate hyperparameter of the Gamma prior distribution of the Weibull shape parameter are used as the first hyperparameter pair, and the shape hyperparameter and prior rate hyperparameter of the Gamma prior distribution of the Weibull scale parameter are used as the second hyperparameter pair. Both the first hyperparameter pair and the second hyperparameter pair are set as joint Jeffreys hyperprior to obtain the second prior layer. Based on the first prior layer and the second prior layer, a two-layer Bayesian prior structure is constructed. In this context, the density function of each joint Jeffreys hyperprior is proportional to the square root of the determinant of the Fisher information matrix of the shape hyperparameter and prior rate hyperparameter of the corresponding Gamma distribution.

3. The method for assessing the theoretical potential of regional wind energy according to claim 2, characterized in that, According to Bayes' theorem, the two-layer Bayesian prior structure is multiplied by the wind speed likelihood function for the target year to generate the joint posterior probability density distribution for each grid, specifically including: Obtain the wind speed likelihood function for the target year, the Gamma prior distribution of the Weibull shape parameter, the Gamma prior distribution of the Weibull scale parameter, the joint Jeffreys hyperprior density of the first hyperparameter pair, and the joint Jeffreys hyperprior density of the second hyperparameter pair. Multiply the wind speed likelihood function of the target year, the Gamma prior distribution of the Weibull shape parameter, the Gamma prior distribution of the Weibull scale parameter, and the joint Jeffreys hyperprior density of the two hyperparameter pairs to obtain the joint posterior probability density function expression of the Weibull parameter and the corresponding Gamma hyperparameter of the target year. The joint posterior probability density function expression is used as the joint posterior probability density distribution for each grid.

4. The method for assessing the theoretical potential of regional wind energy according to claim 1, characterized in that, The process involves transforming the positive parameters of the joint posterior probability density distribution to an unconstrained logarithmic space, sampling using the Hamiltonian Monte Carlo algorithm based on the transformed probability density and corresponding gradient to obtain an initial sample set, performing multi-chain convergence checks and effective sample size evaluations on the initial sample set, and generating a joint posterior sample set for the Weibull shape parameter and the Weibull scale parameter. Specifically, this includes: Set the inter-chain convergence condition and effective sample size threshold, transform the Weibull shape parameter, Weibull scale parameter and the shape hyperparameter and prior rate hyperparameter of the Gamma prior distribution in the joint posterior probability density distribution to the unconstrained logarithmic parameter space to obtain the parameter position vector, and define the potential energy function according to the transformed joint posterior probability density distribution. Among them, the Jacobian correction term for the variable transformation to the unconstrained log-parameter space is automatically included by the posterior sampling procedure when constructing the target probability density of the unconstrained parameter space; An auxiliary momentum variable is introduced into the parameter position vector. A Hamiltonian is constructed through the potential energy function and the auxiliary momentum variable. A dynamic trajectory is generated according to the gradient of the Hamiltonian. Hamiltonian Monte Carlo posterior sampling based on a two-layer Bayesian structure is performed on the dynamic trajectory to obtain an initial sample set. Multiple sampling chains are run in parallel. Based on the Weibull shape parameters and Weibull scale parameters on all sampling chains, as well as the Gamma hyperparameters corresponding to the Weibull shape parameters and Weibull scale parameters, the inter-chain convergence index is calculated. When the inter-chain convergence index of all parameters satisfies the inter-chain convergence condition, the initial sample set of the corresponding spatial grid is determined to have passed the convergence test. The effective sample size is calculated for all parameters that pass the convergence test. When the effective sample size of any parameter to be estimated is lower than the effective sample size threshold, the corresponding grid is determined to have failed the sampling quality test. The posterior sampling is re-executed by increasing the number of preheating times, increasing the number of formal sampling times, or adjusting the sampling control parameters. The samples generated after the preheating period of the sampling chain that passed the quality test are taken as the effective joint posterior sample set.

5. The method for assessing the theoretical potential of regional wind energy according to claim 1, characterized in that, The step of calculating air density and calculating effective operating hours based on the temperature data, air pressure data, digital elevation model data, wind turbine parameter data, and target year wind speed data in the dataset specifically includes: Obtain the hub height, blade radius, cut-in wind speed, and cut-out wind speed from the wind turbine parameter data; The rotor sweep lower edge height and rotor sweep upper edge height are determined based on the hub height and the blade radius, and the vertical interval from the rotor sweep lower edge height to the rotor sweep upper edge height is taken as the rotor sweep vertical interval. The vertical sweeping range of the rotor is divided into multiple micro-element strips, and the center height of each micro-element strip is determined. For each micro-element strip, the air density at the center height of the strip is calculated using the air temperature data, the air pressure data, and the digital elevation model data; Based on the wind speed data for the target year, the cumulative number of hours in each grid where the wind speed falls between the cut-in wind speed and the cut-out wind speed is calculated to obtain the effective operating hours.

6. The method for assessing the theoretical potential of regional wind energy according to claim 5, characterized in that, The step of obtaining the posterior sample set of the regional theoretical wind energy potential through rotor vertical strip integration based on the joint posterior sample set specifically includes: Substitute the Weibull shape parameters and Weibull scale parameters of each group in the joint posterior sample set into the Weibull probability density function, and use the wind shear index to correct the wind speed data in the dataset that has been corrected to the hub height to the center height of each micro-element strip. Combine the air density, the effective operating hours, and the cross-sectional area of ​​each micro-element strip, integrate the wind speed interval from the cut-in wind speed to the cut-out wind speed, and accumulate all micro-element strips to generate the theoretical wind energy potential estimate corresponding to the sample. The theoretical wind energy potential estimates corresponding to all samples are used together as the posterior sample set of the regional theoretical wind energy potential.

7. The method for assessing the theoretical potential of regional wind energy according to claim 1, characterized in that, The calculation of the posterior mean, P90 potential, and uncertainty level of each grid based on the posterior sample set yields the regional wind energy theoretical potential assessment result for the area to be assessed, specifically including: The 10th percentile of each grid is extracted from the posterior sample set as the P90 potential; Based on the posterior sample set, calculate the posterior mean and 95% confidence interval for each grid, and calculate the difference between the upper and lower limits of the 95% confidence interval. The interval width is obtained based on the difference between the upper and lower limits of the 95% confidence interval. The interval width is then divided into low uncertainty level, medium uncertainty level and high uncertainty level according to a preset classification threshold. The uncertainty level corresponding to each grid is obtained based on the classification result. The posterior mean, the P90 potential, and the uncertainty level corresponding to each grid are used as the regional wind energy theoretical potential assessment results for the area to be evaluated and output.

8. A regional wind energy theoretical potential assessment system, characterized in that, The regional wind energy theoretical potential assessment system includes: The multi-source data acquisition module is used to acquire multi-year historical wind speed data, target year wind speed data, temperature data, air pressure data, digital elevation model data, and wind turbine parameter data of the area to be evaluated. Based on the hub height in the wind turbine parameter data, the wind speed height is corrected to the wind turbine hub height, and the datasets of each grid are generated. The wind speed distribution fitting module is used to estimate the Weibull shape parameter and Weibull scale parameter of each grid by year based on the centralized and corrected multi-year historical wind speed data in the dataset, and obtain the historical shape parameter sequence and historical scale parameter sequence. The prior structure construction module is used to construct a Gamma prior distribution of shape parameters based on the historical shape parameter sequence, construct a Gamma prior distribution of scale parameters based on the historical scale parameter sequence, obtain a first prior layer, and use the Jeffreys prior of Gamma hyperparameter as a second prior layer to construct a two-layer Bayesian prior structure. The posterior probability density calculation module is used to establish the target year wind speed likelihood function for each grid based on the centralized and corrected multi-year historical wind speed data of the dataset. According to Bayes' theorem, the two-layer Bayesian prior structure is multiplied with the target year wind speed likelihood function to generate the joint posterior probability density distribution for each grid. The joint posterior sampling module is used to transform the positive parameters of the joint posterior probability density distribution to an unconstrained logarithmic space. Based on the transformed probability density and the corresponding gradient, the Hamilton Monte Carlo algorithm is used to sample to obtain an initial sample set. The initial sample set is then subjected to a multi-chain convergence test and an effective sample size evaluation to generate a joint posterior sample set of Weibull shape parameters and Weibull scale parameters. The wind energy potential calculation module is used to perform air density calculation and effective operating hours statistics based on the temperature data, air pressure data, digital elevation model data, wind turbine parameter data and target year wind speed data in the dataset, and to obtain the posterior sample set of the regional theoretical wind energy potential by rotor vertical strip integration based on the joint posterior sample set. The evaluation and output module is used to calculate the posterior mean, P90 potential, and uncertainty level of each grid based on the posterior sample set, so as to obtain the regional wind energy theoretical potential evaluation result of the area to be evaluated.

9. A terminal, characterized in that, The terminal includes: a memory, a processor, and a regional wind energy theoretical potential assessment program stored in the memory and executable on the processor, wherein when the regional wind energy theoretical potential assessment program is executed by the processor, it implements the steps of the regional wind energy theoretical potential assessment method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a regional wind energy theoretical potential assessment program, which, when executed by a processor, implements the steps of the regional wind energy theoretical potential assessment method as described in any one of claims 1-7.

Images