Wind energy resource endowment assessment method, system, equipment and medium

Abstract

The invention discloses a wind energy resource endowment assessment method, system, device and medium, and relates to the technical field of artificial intelligence and smart meteorology, and the method comprises the steps: obtaining the calculation data of a to-be-assessed region in a target assessment year, and then analyzing the wind speed data and topographic data; constructing a hierarchical feature fusion model and a feature differentiation processing model based on the preprocessed back calculation data to obtain wind energy resource refined feature data; based on the preprocessed reanalysis wind speed data and topographic data, constructing a topographic wind energy generator, and optimizing wind energy resource evaluation to obtain a simulated wind energy resource image; and inputting the wind energy resource refined feature data and the simulated wind energy resource image into a discriminator to obtain a wind energy resource refined evaluation result. According to the method, deep learning and refined feature capture technologies are adopted, refined evaluation of the wind energy resource endowment is achieved, and a scientific basis is provided for reasonable utilization and optimal configuration of wind energy resources.

Description

Technical Field

[0001] The present invention relates to the technical field of artificial intelligence intelligent meteorology, and particularly to a method, system, device and medium for evaluating wind energy resource endowment. Background Art

[0002] In recent years, with the rapid improvement of computing power and the rapid development of deep learning technology, artificial intelligence has made remarkable progress in fields such as perception, evaluation and prediction. Although traditional statistical methods (such as Weibull probability distribution, etc.), wind power density evaluation indicators, and spatial analysis methods based on geographic information system (GIS) can reveal the wind energy resource endowment of a certain area to a certain extent, these methods have obvious limitations: the evaluation means are too single, lacking the ability to comprehensively consider multiple influencing factors; the advantages of deep learning technology in complex pattern recognition and high-dimensional data processing are not fully utilized, resulting in the evaluation results of regional wind energy resources often lacking accuracy and detail. Therefore, there is an urgent need for a model that can integrate multi-feature data and accurately capture the distribution law of wind energy resources to fill the gap in the refined evaluation of regional wind energy resource endowment and provide strong support for the accurate positioning of wind power generation. Summary of the Invention

[0003] In view of the above existing problems, the present invention is proposed.

[0004] Therefore, the present invention provides a method, system, device and medium for evaluating wind energy resource endowment to solve the problem of the limitations of existing wind energy resource evaluation methods.

[0005] To solve the above technical problems, the present invention provides the following technical solutions: In a first aspect, the present invention provides a method for evaluating wind energy resource endowment, including: Obtaining the back-calculated data, reanalysis wind speed data and terrain data of the area to be evaluated in the target evaluation year; Based on the preprocessed back-calculated data, constructing a hierarchical feature fusion model and a feature differentiation processing model, capturing multi-scale features through dilated convolution, adjusting weights by an attention mechanism, and performing convolution and feature conversion respectively after setting thresholds to divide features, to obtain refined wind energy resource feature data; Based on the preprocessed reanalysis wind speed data and terrain data, constructing a terrain wind energy generator to optimize the evaluation of wind energy resources and obtain a simulated wind energy resource image; Inputting the refined wind energy resource feature data and the simulated wind energy resource image into a discriminator to obtain a refined evaluation result of wind energy resources.

[0006] As a preferred solution of a method for evaluating wind energy resource endowment according to the present invention, wherein: The preprocessed back-calculated data includes the following steps: The back-calculated data includes wind speed data, wind power density data, wind shear data, and turbulence intensity data; Normalize the wind speed data and wind power density data in the back-calculated data; Perform conditional normalization on the wind shear data in the back-calculated data; Perform inverse normalization on the turbulence intensity data in the back-calculated data; Stack the normalized wind speed data and wind power density data in the back-calculated data along the channel dimension to obtain the wind energy intensity feature; Stack the conditionally normalized wind shear data and the inverse-normalized turbulence intensity data in the back-calculated data along the channel dimension to obtain the stability feature.

[0007] As a preferred solution of a wind energy resource endowment evaluation method according to the present invention, wherein: The hierarchical feature fusion model includes the following steps: Concatenate the wind energy intensity feature and the stability feature along the time dimension to obtain an original combined feature map; Apply dilated convolution operations with different dilation rates to the combined feature map in the spatial dimension to obtain multi-scale spatial feature maps; Apply dilated convolution operations to the spatial feature maps in the time dimension to capture the changing features in the time series and obtain time feature maps; Combine the spatial feature maps and the time feature maps with the original combined feature map through element-wise multiplication to obtain a fused feature map; Based on the fused feature map, guide the attention mechanism to adjust the weights of the intensity feature and the stability feature to obtain an attention feature map; Integrate the obtained spatial feature maps, time feature maps, and attention feature maps to obtain a comprehensive wind energy resource feature; The application of dilated convolution operations with different dilation rates to the combined feature map in the spatial dimension is expressed as: , wherein, represents the feature map at the t-th time step of and represent the position of the current convolution operation in , and represent the position indices of the convolution kernel, and take values ranging from -1 to 1, represents the receptive field around the position at the current dilation rate; represents the weight in the convolution kernel, Represents the bias term of the convolution operation, Represents the output eigenvalue of the dilated convolution operation at the position Output feature value at the position Represents the spatial dilation rate; The dilated convolution operation on the spatial feature map in the time dimension Is expressed as: , Wherein, Represents the position At the time step Feature at Represents the feature map obtained by spatial dilated convolution, Represents the index in the time dimension, The value range of is from -1 to 1; Represents the time convolution kernel weight, Represents the bias term of the convolution operation, Represents the time dilation rate.

[0008] As a preferred solution of a wind energy resource endowment evaluation method described in the present invention, wherein: The feature differentiation processing model includes the following steps: Adopt average pooling operation to calculate the average values of the wind energy intensity feature and the stability feature in the time dimension respectively, and use the average values as the feature weight thresholds; According to the feature weight thresholds, divide the wind energy intensity feature into relatively strong features and relatively weak features, and divide the wind energy stability feature into relatively stable features and relatively unstable features; Merge the relatively strong features and the relatively stable features to generate enriched features, and apply convolution operation to obtain an enriched feature map; Merge the relatively weak features and the relatively unstable features, and input them into the feature conversion unit to obtain a rich semantic feature map; Merge the enriched feature map and the rich semantic feature map to obtain the differential wind energy resource feature.

[0009] As a preferred solution of a wind energy resource endowment evaluation method described in the present invention, wherein: The preprocessed reanalysis wind speed data and terrain data include the following steps: Resample the reanalysis wind speed data and terrain data to the same spatial resolution as the back-calculated data; Calculate the reanalysis wind power density data using the wind speed data; Normalize the reanalysis wind power density data and terrain data respectively to obtain the normalized reanalysis wind power density data and terrain data.

[0010] As a preferred solution of a method for evaluating wind energy resource endowment according to the present invention, wherein: The steps for constructing the terrain wind energy generator include the following: Merge the normalized reanalysis wind power density data P and the terrain data D, and input them into the terrain wind energy generator; The terrain wind energy generator includes an encoder, a bottleneck layer, and a decoder; Based on the merged input data, the encoder uses a convolutional layer with instance normalization and an activation function to extract key terrain features ; Based on the key terrain features extracted by the encoder , the bottleneck layer uses dilated convolution to capture multi-scale spatial dependencies and generate a high-level feature representation B; Based on the high-level feature representation B provided by the bottleneck layer, the decoder restores the spatial dimension of the image through transposed convolution operations; The decoder combines the key terrain features extracted by the encoder with the normalized reanalysis wind power density data P to generate a simulated wind energy resource image .

[0011] As a preferred solution of a method for evaluating wind energy resource endowment according to the present invention, wherein: The steps for inputting the refined wind energy resource feature data and the simulated wind energy resource image into the discriminator include the following: Integrate the comprehensive wind energy resource features and the differential wind energy resource features to form refined wind energy resource feature data; Use the normalized reanalysis wind power density data to evaluate the accuracy of the refined wind energy resource feature data in restoring local meteorological features; Correct the output of the generator by distinguishing the refined wind energy resource feature data from the simulated wind energy resource image.

[0012] In a second aspect, the present invention provides a wind energy resource endowment evaluation system, including: A data acquisition and preprocessing module, configured to acquire the back-calculated data, reanalysis wind speed data, and terrain data of the area to be evaluated in the target evaluation year; A hierarchical feature fusion and differential processing module, configured to construct a hierarchical feature fusion model and a feature differential processing model based on the preprocessed back-calculated data, capture multi-scale features through dilated convolution, adjust weights through an attention mechanism, and perform convolution and feature transformation respectively after setting thresholds to divide features, so as to obtain refined wind energy resource feature data; A topographic wind energy generator module, which is used to construct a topographic wind energy generator based on preprocessed reanalysis wind speed data and topographic data, optimize the assessment of wind energy resources, and obtain a simulated wind energy resource image; A discriminator module, which is used to input the refined feature data of wind energy resources and the simulated wind energy resource image into the discriminator to obtain a refined assessment result of wind energy resources.

[0013] In a third aspect, the present invention provides an electronic device, including: A memory for storing programs; A processor for executing the computer-executable instructions, and when the computer-executable instructions are executed by the processor, the steps of the method for assessing wind energy resource endowment are implemented.

[0014] In a fourth aspect, the present invention provides a computer-readable storage medium, including: when the program is executed by the processor, the steps of the method for assessing wind energy resource endowment are implemented.

[0015] The beneficial effects of the present invention: By designing a hierarchical feature fusion and feature differentiation processing module, the present invention maximally retains the intensity and stability features of wind energy resources, reduces redundant information, thereby improving the accuracy of wind energy resource assessment and achieving a higher spatial resolution. At the same time, the topographic wind energy generator module generates a wind energy resource image that better conforms to local meteorological characteristics in combination with topographic data, and the discriminator module significantly improves the reliability and accuracy of the assessment model by further optimizing the generation result. Description of the Drawings

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention, and those of ordinary skill in the art can also obtain other drawings based on these drawings without creative efforts.

[0017] Figure 1 It is a schematic diagram of the basic process of a method for assessing wind energy resource endowment provided by an embodiment of the present invention; Figure 2 It is an overall structural diagram of the network model of a method for assessing wind energy resource endowment provided by an embodiment of the present invention; Figure 3 It is a structural diagram of the hierarchical feature fusion module of a method for assessing wind energy resource endowment provided by an embodiment of the present invention; Figure 4 It is a structural diagram of the feature differentiation processing module of a method for assessing wind energy resource endowment provided by an embodiment of the present invention; Figure 5Structural diagram of the feature conversion unit of a wind energy resource endowment assessment method provided by an embodiment of the present invention; Figure 6 Structural diagram of the terrain wind energy generator module of a wind energy resource endowment assessment method provided by an embodiment of the present invention. Detailed implementation manners

[0018] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following will describe the detailed implementation manners of the present invention with reference to the accompanying drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

[0019] Example 1, referring to Figures 1 to 6 , which is an embodiment of the present invention, provides a wind energy resource endowment assessment method, including: S1: Obtain the reanalysis data, reanalysis wind speed data, and terrain data of the area to be evaluated in the target evaluation year; In the embodiment of the present invention, the reanalysis data is the reanalysis data obtained by running the Weather Research and Forecasting model (WRF), and the reanalysis wind speed data is the reanalysis wind speed data of the fifth version of the European Centre for Medium-Range Weather Forecasts Reanalysis (ERA5). The WRF reanalysis data is generated by running the Weather Research and Forecasting model (WRF), and these data are obtained through numerical simulation based on historical meteorological observations (such as global weather stations, satellites, and radar data) and reanalysis data (such as ERA5 and the National Centers for Environmental Prediction NCEP in the United States).

[0020] In the embodiment of the present invention, the WRF reanalysis data, ERA5 reanalysis wind speed data, and terrain data have the following characteristics: The spatial resolution of the WRF reanalysis data is 1 km, and the time resolution is 1 hour, including four elements: wind speed, wind power density, wind shear, and turbulence intensity. To match the analysis requirements, the 1-hour data will be aggregated and converted into daily average data. The spatial resolution of the ERA5 reanalysis wind speed data is 25 km, and the time resolution is also 1 hour. The 1-hour data also needs to be aggregated into daily average data. The spatial resolution of the terrain data is 30 m, covering five elements: altitude, slope, slope curvature, aspect, and flow direction.

[0021] S2: Based on the preprocessed reanalysis data, construct a hierarchical feature fusion model and a feature differentiation processing model, capture multi-scale features through dilated convolution, adjust weights through the attention mechanism, and perform convolution and feature conversion respectively after setting thresholds to divide features, to obtain refined wind energy resource feature data; In the embodiments of the present invention, a traditional statistical analysis method is used to study the long-term variation law of wind energy resources in the area to be evaluated, and the results are used as prior information. Then, the wind speed data (normalized), wind power density data (normalized), wind shear data (conditionally normalized), and turbulence intensity data (inverse normalized) in the WRF back-calculated data are processed. The processed data are stacked pairwise according to the channel dimension to obtain the wind energy intensity feature and stability feature, which are used as the inputs of the hierarchical feature fusion module and the feature differentiation processing module.

[0022] In the embodiments of the present invention, the conditional normalization is specifically as follows: multiple ranges of absolute wind shear values are set, and the probability distribution of historical wind shear data in each range is estimated based on the Weibull probability distribution. The probability distribution of wind energy resources in different ranges is visually analyzed. By comparing the resource differences in each range, the best range of absolute wind shear values that can significantly reflect the potential of wind energy resources in the grids of the area to be evaluated is determined and defined as the best range. According to the best range, the wind shear values during the time period of the wind energy resource endowment to be evaluated are normalized, that is, the wind shear values within this range are close to 1 after normalization, while the wind shear values far from this range will gradually decrease according to the exponential decay rule.

[0023] In the embodiments of the present invention, a hierarchical feature fusion module and a feature differentiation processing module are constructed to optimize the multi-feature information fusion. The hierarchical feature fusion module is used to enhance the spatio-temporal correlation between features, and the feature differentiation processing and aggregation module is used for feature reconstruction and transformation, aiming to reduce the redundant features generated in the feature fusion process and retain more information about small target grids. The outputs of the two modules are integrated as the input of the discriminator module.

[0024] In the embodiments of the present invention, the hierarchical feature fusion module (Hierarchical Feature FusionModule, HFFM), as Figure 3 shown, the specific process is as follows: After splicing the wind energy intensity feature and the stability feature, dilated convolution is used to capture the change patterns of wind energy resources at different scales. The output after the convolution operation is non-linearly processed through the Sigmoid activation function, and the activated result is used to weight the feature map to highlight the most relevant feature parts. Finally, through hierarchical feature fusion, the initial intensity feature, stability feature, and the feature after dilated convolution processing are combined to perform deep feature learning while retaining the original information, so as to more accurately capture the change law of wind energy resources. The specific operations of HFFM are as follows: Wind energy intensity feature and stability feature , for each time step of the feature map Represent different feature states, is the height of the feature map, is the width of the feature map. After feature concatenation, the feature map is obtained. The specific formula is: , where concat is a concatenation operation used to concatenate two or more tensors along a specified dimension. The concatenated feature map is subjected to dilated convolution. The specific operation is as follows: For any position in the feature map, a dilated convolution operation is applied using a 3×3 convolutional kernel and a dilation rate r to extract multi-scale spatial features. At each time step the output can be expressed as: , where is the feature map at the t-th time step of and are the positions of the current convolution operation in , and are the position indices of the convolutional kernel, traversing from -1 to 1 respectively; represents the receptive field around the position at the current dilation rate; is the weight in the convolutional kernel; is the bias term of the convolution operation, is the output feature value of the dilated convolution operation at the position . The entire convolution process is to continuously perform convolution operations on the local regions of each position in each time step, and finally concatenate to obtain the spatial feature map . . represents the spatial dilation rate. For different dilation rates: when = 1, the convolution output is ; when = 2, the convolution output is ; when = 3, the convolution output is .

[0025] Perform dilated convolution operation on the feature map after spatial convolution in the time dimension . For each time step of any position in the feature map, perform convolution with a dilation rate . The output is expressed as: , Among them, is the position at the time step where is the feature map obtained through spatial dilated convolution, is the index in the time dimension, ranging from -1 to 1; is the time convolution kernel weight; is the bias term of the convolution operation. The entire convolution process is to continuously perform convolution operations with different receptive fields on each time step at each position in , and finally splice to obtain the time feature map . represents the time dilation rate. For different dilation rates: when = 1, the convolution output is ; when = 2, the convolution output is ; when = 3, the convolution output is . Multiply the outputs of spatial and temporal dilated convolutions element-wise with to obtain the fused feature , enabling the model to "strengthen" or "adjust" important features at specific positions and time steps.

[0026] Fuse features using the attention mechanism , and , aiming to use the high-level feature to guide the attention mechanism to more efficiently adjust the fusion weights of the wind energy intensity feature and the stability feature , automatically focus on more important features or regions, ignore less relevant parts, and obtain more representative wind energy resource endowment features.

[0027] In the embodiment of the present invention, the feature differentiation processing module (Feature Differentiation Processing Module, FDPM), as shown in Figure 4 , the specific process is as follows: The average pooling operation is used to calculate the average values of the wind energy intensity feature and the stability feature in the time dimension respectively, and the average values are used as the feature weight thresholds, so as to capture the global information of each spatial position at all time steps. According to the feature weight thresholds, the wind energy intensity feature is divided into relatively strong features and relatively weak features, and the wind energy stability feature is divided into relatively stable features and relatively unstable features. Then, the relatively strong features and the relatively stable features are merged to generate enriched features, and convolutional operations are applied to extract more detailed information; the relatively weak features and the relatively unstable features are merged and input into the feature transformation unit to generate feature maps with richer semantic information using fewer computing resources. Finally, feature merging is performed to obtain more detailed wind energy resource endowment features after FDPM processing. The specific operations of FDPM are as follows: Wind energy intensity feature And stability feature , each time step Of the feature map Represents different feature states, Is the height of the feature map, Is the width of the feature map. The average pooling operation is used to calculate the wind energy intensity feature And stability feature The average value in the time dimension to obtain the feature weight threshold. This operation captures the global information of each spatial position at all time steps. The specific formula is: , , Among them, Is The Th time step of the feature map, Is the weight threshold of the wind energy intensity feature, Is The Th time step of the feature map, Is the weight threshold of the wind energy stability feature. According to the calculated feature weight thresholds, the wind energy intensity feature Is divided into relatively strong features and relatively weak features; the wind energy stability feature Is divided into relatively stable features and relatively unstable features. The specific formula is: , , , , Among them, Is the relatively strong feature, Is the relatively weak feature, Is a relatively stable feature, Is a relatively unstable feature. Since there are differences in the time steps after the features are classified by the threshold, the relatively strong features and the relatively stable features are added element by element at the same time step, while the features at different times are not added. The features generated after this processing are enriched features . Further extract more detailed information through convolution operations to enhance the expression ability of the features. The convolution operations are performed on each position of the feature map at the time step : , Among them, And Are the positions at the -th time step of the enriched feature for the current convolution operation; And Are the position indices of the convolution kernel, traversing from 1 to 3 respectively; Indicates the spatial convolution range of the convolution operation at the position ; Is the weight in the convolution kernel; Is the bias term of the convolution operation; Is the value at the position In the -th feature map of the convolution output.

[0028] Add the relatively weak features and the relatively unstable features element by element at the same time step, while the features at different times are not added. The features generated after this processing Are input to the Feature Transformation Unit (FTU) to obtain , which uses fewer computing resources to generate a feature map with richer semantic information.

[0029] Add the output result after convolving the enriched feature And Element by element to obtain more detailed wind energy resource endowment features after FDPM processing .

[0030] In the embodiments of the present invention, as Figure 5 Shown, the specific operations of the FTU are as follows: Use depthwise separable convolution to extract Local features, and the specific formula is: , Among them, Represents half of the size of the depth convolution kernel (for example, a 3×3 convolution corresponds to ); Represents the weights of the depth convolution kernel (usually of size ); Represents of the value at position in the feature map at time ; Represents of the output after depth convolution at position in the feature map at time . Then, a 1×1 convolution is used to perform a linear combination of all the time channels of the depth convolution output, thereby achieving information fusion between channels: , where represents the weight matrix of the pointwise convolution, which is responsible for converting input channels into output channels; represents the feature value at position in the feature map at time ; represents the fusion result of all the time channels at position .

[0031] Generate attention weights through dimensionality reduction by the pooling layer, feature extraction by 3×3 convolution operation, and normalization by the normalization layer, for adaptively controlling the importance of local features. The specific formula is: , where represents the pooling operation; represents the convolution kernel weights; represents the convolution operation; and represent the mean and standard deviation of the normalization respectively; and represent the trainable parameters of the normalization; is a small number to prevent the denominator from being zero; represents the sigmoid activation function; is used as the attention weight for feature enhancement. Multiply element-wise with to enhance the feature expression of important regions, and the output result is .

[0032] S3: Based on the preprocessed reanalysis wind speed data and terrain data, construct a terrain wind energy generator, optimize the wind energy resource assessment, and obtain a simulated wind energy resource image; In the embodiments of the present invention, a topographic wind energy generator module is constructed, adopting an architecture inspired by U-Net and optimized for wind energy resource assessment tasks. This generator module uses topographic data to replace the noise input in traditional cGAN and combines wind power density data to generate simulated wind energy resource images.

[0033] In the embodiments of the present invention, the topographic wind energy generator module adopts an architecture inspired by U-Net and is modified for refined wind energy resource tasks. The encoder uses convolutional layers with instance normalization and LeakyReLU activation to compress the spatial dimension while increasing the feature depth. The bottleneck layer uses dilated convolutions to capture multi-scale spatial dependencies, which is crucial for understanding wind energy patterns of different intensities within the area to be evaluated. The decoder uses transposed convolutions and skip connections to improve the quality of wind energy resource assessment while retaining spatial details.

[0034] In the embodiments of the present invention, as Figure 6 shown, the specific operations of the topographic wind energy generator module are as follows: The topographic data and wind power density data are merged to form a new input, and the specific formula is: , = , where, represents the topographic data, which is the main input of the generator. The generator extracts features from the topographic data through the encoder and then generates an output related to the wind power density; represents the reanalyzed wind power density data after normalization, representing the wind energy density at each location under different time steps, and is used to conditionally constrain the generator to generate a matching wind energy image. The convolutional layer extracts the spatial features of the channel of the input data through different filters and strides and introduces non-linear characteristics using the LeakyReLU activation function; represents the padding in the convolution operation; represents the bias term in the convolutional layer of the encoder; the output is the high-dimensional representation of the topographic features and will be used as the input of the bottleneck layer. The bottleneck layer uses dilated convolutions to capture spatial dependencies at different scales, which is crucial for understanding the spatial patterns of wind energy distribution. The specific formula is: , where, is the kernel size, is the dilation rate, represents the bias term in the convolutional layer of the bottleneck layer. The dilated convolution ( It can enable the generator to capture the spatial features of wind energy over a larger range. The decoder gradually restores the spatial resolution and combines the terrain features extracted by the encoder with the wind power density data through skip connections. The specific formula is as follows: , , where, represents the bias term in the transposed convolution layer of the decoder. The transposed convolution ( ) operation restores the spatial dimension of the image through the transposed convolution operation to generate the wind energy resource image . The skip connection combines the terrain features from the encoder and the wind power density data into the input of the decoder, helping the terrain wind energy generator to better consider the relationship between terrain and wind energy resources when generating the wind energy resource map, so as to generate a wind energy resource map that can accurately reflect the spatial distribution of wind energy potential and is closely combined with the actual terrain characteristics .

[0035] S4: Input the refined feature data of wind energy resources and the simulated wind energy resource image into the discriminator to obtain the refined evaluation result of wind energy resources.

[0036] In the embodiment of the present invention, a discriminator module is constructed to receive the refined feature data of wind energy resources from S2, the reanalysis wind power density data in S3, and the simulated wind energy resource image. As the "quality controller" in the RWEA-MFFC model, the discriminator module can not only use the wind power density data to check the effect of the refined feature data of wind energy resources in restoring local meteorological features, but also correct the output of the generator module by distinguishing the refined feature data of wind energy resources from the generated data, so that it gradually approaches the real wind energy resource features, thereby improving the reliability and accuracy of the evaluation result.

[0037] In the embodiment of the present invention, the training and verification process of the model is also described, including: Select the year of the wind energy resource endowment to be evaluated (such as 2023 or 2005 - 2009), and use the WRF back-calculated data, ERA5 reanalysis wind speed data, and terrain data during that year as the model input data to build the training dataset of the RWEA-MFFC model.

[0038] According to the built training dataset, calculate the loss of the RWEA-MFFC model, optimize the network parameters, and obtain a network model that conforms to the RWEA-MFFC architecture.

[0039] Compare the grids with relatively high wind energy resource potential in the refined wind energy resource assessment results output by the RWEA-MFFC model with the site data of meteorological observation stations in the area to be evaluated, and verify the ability of the model to restore the wind energy resource endowment of local grids.

[0040] It should be noted that the present invention discloses a method for refined wind energy resource endowment assessment (Refined Wind Energy Resource Endowment Assessment Based on Multi-Feature Fusion and cGAN, RWEA-MFFC), which can not only verify the consistency of the multi-index wind energy resource assessment results based on WRF back-calculated data with ERA5 data in the spatio-temporal distribution of wind energy resources, but also restore the local wind energy resource characteristics lost due to the low resolution of ERA5 data, so as to achieve refined assessment of the wind energy resource endowment of the area to be evaluated.

[0041] Embodiment 2, which is an embodiment of the present invention, provides a method for wind energy resource endowment assessment. In order to verify its beneficial effects, scientific demonstration is carried out through specific implementation methods and implementation effects.

[0042] The specific implementation of this embodiment is as follows: The experimental results shown in this embodiment rely on the research area between 24.51°N and 29.29°N, 103.49°E and 110.12°E. This area roughly covers the main part of a certain province in China, including complex terrain features and diverse climate conditions. The selection of this area depends on its unique geographical and meteorological characteristics, making it an ideal place for the research of wind energy resource endowment assessment models. The experimental results show that the method of the present invention has high accuracy and reliability in back-calculating and comprehensively evaluating historical wind energy resource data in this area.

[0043] This embodiment uses the data from 2000 to 2004, 2006 to 2010, 2012 to 2016, and 2018 to 2022 as the training set, and the data from 2005, 2011, 2017, and 2023 as the test set.

[0044] In this embodiment, through the analysis of the existing ERA5 wind power density data, it is found that the data lacks refinement in spatial distribution and fails to effectively distinguish between rich and poor wind energy resource regions. In addition, when comparing with the actual site data of a certain provincial meteorological observation station, it is found that the ERA5 data has an obvious underestimation in the estimation of wind energy resources and fails to accurately reflect the true differences in wind energy resources within a certain province. In contrast, the refined wind energy resource endowment assessment method based on multi-feature fusion and conditional generative adversarial network proposed in the present invention can better fit the site data, and shows an obvious similar trend with the ERA5 data in the spatio-temporal distribution of wind energy resource endowment, while verifying that the method of the present invention can recover the loss of local wind energy resource characteristics caused by the low resolution of ERA5.

[0045] To further verify the advantages of the method of the present invention in the assessment of wind energy resource endowment in a certain province, a module ablation experiment was carried out. The experimental results show that the wind energy resource distribution map generated by the traditional generative adversarial network method is similar to the ERA5 data and both fail to fully reflect the spatial differences in wind energy resource endowment in a certain province. The wind energy resource distribution map after adding HFFM to the generative adversarial network method can show certain regional wind energy resource differences, but still fails to accurately fit the site data. In the wind energy resource distribution map presented after adding FDPM to the generative adversarial network method, although this module effectively recovers the underestimation in the ERA5 data by learning the wind energy stability characteristics, it fails to accurately present the differences in wind energy resource endowment between regions. The combined wind energy resource distribution map of HFFM and FDPM is significantly better than using either module alone compared to using them separately. Further replacing the generator of the generative adversarial network with the terrain wind energy generator module, the presented wind energy resource distribution map can better fit the site data and shows a significant similarity with the ERA5 data in the spatio-temporal distribution of wind energy resource endowment.

[0046] Embodiment 3, in this embodiment, a wind energy resource endowment assessment system is provided, including: A data acquisition and preprocessing module, configured to acquire the back-calculated data, reanalysis wind speed data, and terrain data of the area to be evaluated in the target evaluation year; A hierarchical feature fusion and differentiation processing module, configured to construct a hierarchical feature fusion model and a feature differentiation processing model based on the preprocessed back-calculated data to obtain refined wind energy resource feature data; A terrain wind energy generator module, configured to construct a hierarchical feature fusion model and a feature differentiation processing model based on the preprocessed back-calculated data, capture multi-scale features through dilated convolution, adjust weights through an attention mechanism, and perform convolution and feature transformation respectively after setting thresholds to divide features to obtain refined wind energy resource feature data; A discriminator module, which is used to input the refined feature data of wind energy resources and the simulated wind energy resource images into the discriminator to obtain the refined evaluation result of wind energy resources.

[0047] This embodiment also provides an electronic device, which is applicable to a situation of a wind energy resource endowment evaluation method, including: A memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement a wind energy resource endowment evaluation method as proposed in the above embodiment.

[0048] This embodiment also provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, it implements a wind energy resource endowment evaluation method as proposed in the above embodiment.

[0049] The storage medium proposed in this embodiment and the implementation of a wind energy resource endowment evaluation method proposed in the above embodiment belong to the same inventive concept. The technical details not described in detail in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0050] From the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software and necessary general-purpose hardware. Of course, it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, in essence, or the part that makes a contribution to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a floppy disk, a read-only memory (ROM), a random access memory (RAM), a flash memory (FLASH), a hard disk, or an optical disc of a computer, etc., including several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods of various embodiments of the present invention.

[0051] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

Claims

1. A method for evaluating wind energy resource endowment, characterized in that: include: Obtain back-calculated data, reanalyzed wind speed data, and terrain data for the area to be assessed in the target assessment year; Based on the preprocessed back-calculation data, a hierarchical feature fusion model and a feature differentiation processing model are constructed. The multi-scale features are captured by dilated convolution, the weights are adjusted by the attention mechanism, and the convolution and feature conversion are performed after setting the threshold to divide the features, so as to obtain the refined feature data of wind energy resources. Based on the preprocessed reanalyzed wind speed data and terrain data, a terrain wind energy generator is constructed to optimize the wind energy resource assessment and obtain a simulated wind energy resource image; The refined feature data of wind energy resources and simulated wind energy resource images are input into the discriminator to obtain the refined evaluation results of wind energy resources.

2. A wind energy resource endowment assessment method according to claim 1, characterized in that: The preprocessed back-calculation data includes the following steps: The back-calculated data includes wind speed data, wind power density data, wind shear data and turbulence intensity data; Normalize the wind speed data and wind power density data in the back-calculated data; Perform conditional normalization on the wind shear data in the back-calculated data; Denormalize the turbulence intensity data in the back-calculated data; The normalized wind speed data and wind power density data in the back-calculated data are stacked in the channel dimension to obtain the wind energy intensity characteristics; The conditionally normalized wind shear data and the denormalized turbulence intensity data in the back-calculated data are stacked in the channel dimension to obtain the stability characteristics.

3. A wind energy resource endowment assessment method according to claim 2, characterized in that: The hierarchical feature fusion model includes the following steps: The wind energy intensity characteristics and stability characteristics are spliced ​​in the time dimension to obtain the original combined characteristic map; Apply different dilation rates to the combined feature map to perform dilated convolution operations in the spatial dimension to obtain multi-scale spatial feature maps; Perform an expansion convolution operation on the spatial feature map in the time dimension to capture the changing characteristics of the time series and obtain a time feature map; The spatial feature map and the temporal feature map are combined with the original combined feature map by element-wise multiplication to obtain a fused feature map; Based on the fusion feature map, the attention mechanism is guided to adjust the weights of the intensity feature and the stability feature to obtain the attention feature map; Integrate the acquired spatial feature map, temporal feature map and attention feature map to obtain comprehensive wind energy resource characteristics; The dilated convolution operation on the spatial dimension by applying different dilation rates to the combined feature map is expressed as: ， in, express The feature map of the t-th time step, and Indicates that the current convolution operation is The position in and Represents the position index of the convolution kernel, and The value range is from -1 to 1. Indicates the position at the current expansion rate The surrounding receptive field; represents the weight in the convolution kernel, represents the bias term of the convolution operation, Indicates that the dilated convolution operation is at position The output eigenvalue at represents the spatial expansion rate; The spatial feature map is subjected to an expansion convolution operation in the time dimension. It is expressed as: ， in, Indicates location At time step The above characteristics, represents the feature map obtained by spatial dilation convolution, Represents the index on the time dimension, The value range is from -1 to 1; represents the temporal convolution kernel weight, represents the bias term of the convolution operation, represents the time dilation rate.

4. A wind energy resource endowment assessment method according to claim 3, characterized in that: The feature differentiation processing model includes the following steps: The average pooling operation is used to calculate the average values ​​of wind energy intensity and stability characteristics in the time dimension, and the average values ​​are used as the feature weight threshold. According to the feature weight threshold, the wind energy intensity feature is divided into relatively strong features and relatively weak features, and the wind energy stability feature is divided into relatively stable features and relatively unstable features; The relatively strong features and the relatively stable features are combined to generate enriched features, and the convolution operation is applied to obtain the enriched feature map; Combine relatively weak features with relatively unstable features and input them into the feature conversion unit to obtain rich semantic feature mapping; The enriched feature map and the rich semantic feature map are merged to obtain the differential wind energy resource characteristics.

5. A wind energy resource endowment assessment method according to claim 4, characterized in that: The pre-processed reanalysis of wind speed data and terrain data comprises the following steps: The reanalysis wind speed data and topographic data were resampled to the same spatial resolution as the back-calculated data; Reanalysis wind power density data are calculated using wind speed data; The reanalysis wind power density data and terrain data are normalized respectively to obtain normalized reanalysis wind power density data and terrain data.

6. A wind energy resource endowment assessment method according to claim 5, characterized in that: The construction of the terrain wind energy generator comprises the following steps: The normalized reanalysis wind power density data P and terrain data are combined D and input into the terrain wind energy generator; The terrain wind energy generator includes an encoder, a bottleneck layer and a decoder; Based on the merged input data, the encoder uses convolutional layers with instance normalization and activation functions to extract key features of the terrain. ; Key features of terrain extracted based on encoder ,The bottleneck layer uses dilated convolution to capture multi-scale spatial dependencies and generate high-level feature representation B; Based on the high-level feature representation B provided by the bottleneck layer, the decoder restores the spatial dimension of the image through transposed convolution operations; The decoder uses skip connections to extract the key features of the terrain from the encoder. Combined with the normalized reanalysis wind power density data P, a simulated wind energy resource image is generated. .

7. A wind energy resource endowment assessment method according to claim 6, characterized in that: The step of inputting the refined characteristic data of wind energy resources and the simulated wind energy resource image into the discriminator comprises the following steps: Integrate comprehensive wind energy resource characteristics and differential wind energy resource characteristics to form refined wind energy resource characteristic data; The normalized reanalysis wind power density data are used to evaluate the accuracy of the refined wind energy resource characteristic data in restoring local meteorological characteristics. By distinguishing the refined characteristic data of wind energy resources from the simulated wind energy resource images, the output of the generator is corrected.

8. A wind energy resource endowment assessment system, using a wind energy resource endowment assessment method as claimed in any one of claims 1 to 7, characterized in that: include: Data acquisition and preprocessing module, used to obtain back-calculated data, reanalyzed wind speed data and terrain data of the area to be assessed in the target assessment year; Hierarchical feature fusion and differentiation processing module, which is used to build hierarchical feature fusion model and feature differentiation processing model based on preprocessed back-calculation data, capture multi-scale features through dilated convolution, adjust weights through attention mechanism, and perform convolution and feature conversion respectively after setting thresholds to divide features, so as to obtain refined feature data of wind energy resources; The terrain wind energy generator module is used to build a terrain wind energy generator based on the preprocessed reanalyzed wind speed data and terrain data, optimize the wind energy resource assessment, and obtain the simulated wind energy resource image; The discriminator module is used to input the refined feature data of wind energy resources and the simulated wind energy resource images into the discriminator to obtain the refined evaluation results of wind energy resources.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of a wind energy resource endowment assessment method described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of a wind energy resource endowment assessment method according to any one of claims 1 to 7 are implemented.

Images