Regional wind power icing withdrawal capacity clustering set prediction method and device
By clustering and model fusion of wind farms, the problems of turbine model differences and randomness of human intervention in wind power icing reserve capacity prediction in existing technologies have been solved, and more accurate regional wind power icing reserve capacity prediction has been achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-29
- Publication Date
- 2026-04-07
AI Technical Summary
Existing wind power icing de-icing capacity prediction methods mainly use individual wind farms as the basic unit, which cannot meet the needs of provincial power system dispatch for global prediction of icing de-icing capacity. They also suffer from randomness issues caused by turbine model differences and human intervention, making it difficult to achieve accuracy and reliability.
By clustering wind farms in the region based on meteorological and topographical features, the region is divided into multiple wind farm clusters. Individual and benchmark prediction models are established, and the region's total wind power icing reserve capacity is obtained by weighted fusion of the fusion weights and benchmark prediction values.
It improves the accuracy and reliability of regional wind power icing and reserve capacity prediction, eliminates random errors caused by turbine model differences and human intervention, and realizes the transformation from unpredictable strong randomness of wind farms to predictable cluster regularity.
Smart Images
Figure CN121808433A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of new energy technology, specifically to a method, device, storage medium, and computer program product for clustered prediction of regional wind power icing back-off capacity. Background Technology
[0002] As wind power continues to increase its share in the power system, the stability of wind power output is crucial for the safe dispatch of the power grid. Under harsh weather conditions such as low temperatures and high humidity in winter, critical components of wind turbines, such as blades and hubs, are prone to icing, leading to decreased aerodynamic performance, abnormal loads, and even equipment failure. Therefore, it is necessary to ensure operational safety through reserve shutdown (i.e., suspending the operation of some wind turbines). However, large-scale wind power icing and reserve shutdown can significantly impact the advance decision-making of provincial power dispatching departments.
[0003] Currently, wind power icing decommissioning capacity prediction is mainly carried out on a single wind farm as the basic unit. However, this model has significant technical defects and is difficult to meet the needs of provincial power system dispatch for global prediction of icing decommissioning capacity: (1) The difference in turbine models from different turbine manufacturers leads to inconsistent shutdown judgment standards; (2) When predicting the risk of icing, some wind farm operation and maintenance personnel will manually shut down the turbines in advance based on experience to avoid equipment damage caused by increased icing. The timing and scale of manual shutdown have a low degree of matching with weather forecasts, and the operating habits of different operation and maintenance teams are different, which makes the decommissioning capacity of a single farm highly random. Existing prediction models based on farms cannot effectively fit this kind of random fluctuation.
[0004] Existing methods for predicting wind power icing capacity reduction have significant shortcomings when dealing with issues such as turbine model differences and human intervention. Therefore, there is an urgent need for a clustered prediction technology that can adapt to the needs of provincial-level regions to improve the accuracy and reliability of regional wind power capacity reduction prediction. Summary of the Invention
[0005] The purpose of this application is to provide a method, apparatus, storage medium, and computer program product for clustered prediction of regional wind power icing back-off capacity.
[0006] To achieve the above objectives, the first aspect of this application provides a clustered prediction method for regional wind power icing backup capacity, comprising: Based on meteorological and topographical features, wind farms in the region are clustered into multiple wind farm clusters. For each wind farm cluster, an individual icing reserve capacity prediction model is established for each wind farm within the cluster, and a benchmark prediction model for the icing reserve capacity of the wind farm cluster is established based on the historical aggregated data of the wind farm cluster. The individual icing-induced reserve capacity prediction model is used to obtain the individual reserve capacity prediction values for each wind farm within each wind farm cluster. For each wind farm cluster, the fusion weight of each wind farm in the cluster is determined based on the similarity of meteorological and topographic features of each wind farm in the cluster. The fusion weight is then used to weight and fuse the individual reserve capacity prediction values of each wind farm to obtain the preliminary set value of the wind farm cluster. The baseline prediction value of each wind farm cluster is obtained by using the baseline prediction model of each wind farm cluster, and the initial value of the set is fused and corrected with the baseline prediction value to obtain the icing backup capacity prediction value of each wind farm cluster. The total predicted wind power icing reserve capacity of the region is obtained by summing the predicted values of the icing reserve capacity of all wind farm clusters.
[0007] In this embodiment, clustering wind farms within a region into multiple wind farm clusters based on meteorological and topographic features includes: acquiring meteorological and topographic data for each wind farm within a preset historical period; selecting feature parameters for clustering from the meteorological and topographic data, wherein the meteorological feature parameters include at least the average daily minimum temperature, average daily relative humidity, and average daily wind speed in winter, and the topographic feature parameters include at least the average altitude, average slope, and quantified value of landform type; standardizing the selected feature parameters; and using a clustering algorithm to calculate the Euclidean distance between each wind farm based on the standardized feature parameters, and dividing the wind farms within the region into multiple wind farm clusters according to the Euclidean distance.
[0008] In this embodiment of the application, determining the fusion weight of each wind farm based on the similarity of meteorological and topographic feature parameters includes: calculating the average feature similarity between each wind farm and all other wind farms in the cluster; and normalizing the average feature similarity to obtain the fusion weight of each wind farm.
[0009] In this embodiment of the application, the initial value of the set of each wind farm cluster is calculated using the following formula:
[0010] in, This represents the initial set of values for the wind farm cluster. Let be the weight of the j-th wind farm in the wind farm cluster; Let be the predicted individual reserve capacity of the j-th wind farm; M is the total number of wind farms in the wind farm cluster.
[0011] In this embodiment of the application, the predicted icing backup capacity of each wind farm cluster is calculated using the following formula:
[0012] in, Forecast values of icing fallback capacity for wind farm clusters; The fusion coefficient; This represents the initial set of values for the wind farm cluster. For the benchmark of wind farm clusters Measured value.
[0013] In the embodiments of this application, the fusion coefficient The determination was made through optimization using historical validation set data.
[0014] In this embodiment of the application, the historical aggregated data includes at least: time series aggregation of historical icing monitoring data of all wind farms in the cluster, time series aggregation of historical reserve capacity records, and time series of corrected meteorological data covering the area where the cluster is located; the individual icing reserve capacity prediction model is trained based on the following data of the corresponding wind farm: historical icing event data of the wind farm, historical reserve operation record data, and time series of measured meteorological data at the wind farm site.
[0015] A second aspect of this application provides a regional wind power icing backfill capacity clustered prediction device, comprising: The memory is configured to store instructions; The processor is configured to retrieve instructions from memory and, when executing the instructions, to implement the aforementioned regional wind power icing back-off capacity clustered prediction method.
[0016] A third aspect of this application provides a machine-readable storage medium storing instructions that, when executed by a processor, configure the processor to perform the aforementioned regional wind power icing backup capacity clustered prediction method.
[0017] The fourth aspect of this application provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method for clustered prediction of regional wind power icing back-up capacity.
[0018] The technical solution provided in this application is based on the principle that wind fields are influenced by large-scale atmospheric circulation systems and local topographic elements. If the current wind field is similar to historical wind fields and their dynamic evolution is similar, then the output power of the wind farm will be similar. To eliminate similarity errors, this application proposes using ensemble averaging for forecasting, thereby eliminating the randomness of similarity forecasts and making the forecast results more accurate.
[0019] Other features and advantages of the embodiments of this application will be described in detail in the following detailed description section. Attached Figure Description
[0020] The accompanying drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the following detailed description to explain the embodiments of this application, but do not constitute a limitation on the embodiments of this application. In the drawings: Figure 1 The schematic diagram illustrates a process flow of a clustered prediction method for regional wind power icing backup capacity according to an embodiment of this application; Figure 2 The diagram illustrates the internal structure of a computer device according to an embodiment of this application. Detailed Implementation
[0021] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for illustration and explanation of the embodiments of this application and are not intended to limit the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.
[0022] Figure 1 The illustration schematically shows a flowchart of a clustered prediction method for regional wind power icing backup capacity according to an embodiment of this application. Figure 1 As shown in one embodiment of this application, a clustered prediction method for regional wind power icing backup capacity is provided. This embodiment mainly illustrates the application of this method to a processor or controller, and includes the following steps: Step 102: Based on meteorological and topographical features, the wind farms in the region are clustered into multiple wind farm clusters. Step 104: For each wind farm cluster, establish an individual icing reserve capacity prediction model for each wind farm within the wind farm cluster, and establish a benchmark prediction model for the icing reserve capacity of the wind farm cluster based on the historical aggregated data of the wind farm cluster. Step 106: Obtain the individual reserve capacity prediction value of each wind farm in each wind farm cluster through the individual icing reserve capacity prediction model. Step 108: For each wind farm cluster, determine the fusion weight of each wind farm in the cluster based on the similarity of meteorological and topographic feature parameters of each wind farm in the cluster, and use the fusion weight to perform weighted fusion of the individual reserve capacity prediction values of each wind farm to obtain the preliminary set value of the wind farm cluster. Step 110: Obtain the baseline prediction value of each wind farm cluster through the baseline prediction model of each wind farm cluster, and merge and correct the initial value of the set with the baseline prediction value to obtain the icing backup capacity prediction value of each wind farm cluster. Step 112: Sum the predicted values of icing reserve capacity for all wind farm clusters to obtain the total predicted value of regional wind power icing reserve capacity.
[0023] Icing shutdown capacity refers to the generating capacity of a wind turbine that is proactively shut down to ensure equipment safety when components such as turbine blades and hubs become icy due to low temperature and high humidity. Meteorological characteristic parameters are quantitative indicators describing the atmospheric physical state of the area where the wind farm is located. In this scheme, meteorological characteristic parameters are those highly correlated with icing formation, including daily average minimum temperature, daily average relative humidity, and daily average wind speed. Daily average minimum temperature refers to the lowest air temperature each day. Daily average relative humidity refers to the water vapor content in the air. Daily average wind speed refers to wind speed, which affects the morphology and accumulation rate of icing. These parameters collectively determine the meteorological conditions under which icing will occur and at what rate. Topographic characteristic parameters are static spatial indicators describing the geographical environment where the wind farm is located.
[0024] In one embodiment, clustering wind farms within a region into multiple wind farm clusters based on meteorological and topographic features includes: acquiring meteorological and topographic data for each wind farm within a preset historical period; selecting feature parameters for clustering from the meteorological and topographic data, wherein the meteorological feature parameters include at least the average daily minimum temperature, average daily relative humidity, and average daily wind speed in winter, and the topographic feature parameters include at least the average altitude, average slope, and quantified values of landform type; standardizing the selected feature parameters; and using a clustering algorithm to calculate the Euclidean distance between each wind farm based on the standardized feature parameters, and dividing the wind farms within the region into multiple wind farm clusters according to the Euclidean distance.
[0025] In this scheme, "ice accumulation" can refer to characteristics that can influence icing distribution by altering local microclimates, including average altitude, average slope, and quantified landform type. Higher altitudes generally have lower temperatures, making icing more likely. Topographic slope affects airflow lifting and cooling processes. Quantified landform type refers to converting landform categories such as mountains, hills, and plains into numerical codes, such as 1, 2, 3, etc., for computer processing. First, the processor can collect historical meteorological and topographic data for all wind farms in the region, such as the average daily minimum temperature, average daily relative humidity, and average daily wind speed for the past three winter months (December, January, and February). Topographic feature parameters include the wind farm's average altitude, average slope, and landform type, such as mountains, hills, and plains, quantified into numerical codes. All feature parameters are normalized to eliminate the influence of dimensional differences on the clustering results; the standardization formula is: , Where x i These are the original eigenvalues. For the standardized value, min(x) i ), max(x i ) are the minimum and maximum values of the feature parameter, respectively.
[0026] Since icing risk is primarily controlled by local weather and topography, it's crucial to first identify wind farms with similar environments and divide them into several groups (clusters). This makes the icing problem within each group relatively homogeneous and predictable. Then, wind farms within the region can be clustered into multiple wind farm clusters, ensuring that wind farms within the same cluster have highly similar icing environments (weather and topography). Specifically, a clustering algorithm is used to cluster all wind farms within a provincial area. During the clustering process, using wind farms as samples and standardized feature parameters as dimensions, the Euclidean distance between samples is calculated, grouping wind farms that are close together into the same cluster. This ultimately yields N wind farm clusters with similar weather and topographical conditions. For each cluster identified in the previous step, two types of prediction models are established in parallel: individual prediction models and baseline prediction models, namely, individual icing reserve capacity prediction models and icing reserve capacity baseline prediction models. The individual icing reserve capacity prediction model is a dedicated model for each wind farm within the cluster. The icing reserve capacity baseline prediction model is a unified model for the entire cluster. This model learns the commonalities of the cluster, such as the average reserve capacity reduction pattern across the entire region.
[0027] Specifically, the individual icing-induced reserve capacity prediction model is a prediction model built for a single, specific wind farm. This model establishes its personalized weather-reservation response relationship by learning from the wind farm's own historical data. Its prediction results reflect the expected reserve behavior of the wind farm under specific weather conditions, but may include random errors due to differences in turbine models and human operating habits. Through an independent model for each wind farm, the individual reserve capacity prediction value C for each member is obtained. j (j=1,2,...,M, where M is the number of wind farms in the cluster).
[0028] Historical aggregated data for wind farm clusters refers to a dataset formed by summarizing and merging historical data from all wind farms within a cluster. For example, it might involve merging and statistically analyzing icing records from all wind farms within a cluster, or averaging and correcting regional meteorological model data across the cluster. Historical aggregated data represents the collective memory and average state of the entire cluster. The icing reserve capacity baseline prediction model is a prediction model built for the entire wind farm cluster. This model is trained based on the aforementioned historical aggregated data, and its output reflects the expected average reserve capacity level under the average meteorological and topographical conditions of the cluster. Individual icing reserve capacity prediction models can obtain individual reserve capacity prediction values for each wind farm within each wind farm cluster, essentially collecting preliminary judgments from all members within the cluster.
[0029] Since the voices of each wind farm should not be equal, the opinions of members with higher feature similarity to the overall cluster environment should be given more weight. Therefore, the system calculates the fusion weight of each wind farm based on similarity, and then uses these weights to weight and fuse all individual prediction values to obtain a preliminary set value. This process simulates expert consultation, giving greater weight to more representative opinions, thereby smoothing out prediction noise caused by special reasons of individual wind farms, such as accidental human intervention. Feature similarity is used to quantify the degree of similarity between any two wind farms in meteorological and topographic features within the same cluster. The higher the similarity, the more likely they are to perform consistently in terms of icing risk. Specifically, in one embodiment, determining the fusion weight of each wind farm based on the similarity of meteorological and topographic feature parameters includes: calculating the average feature similarity between each wind farm and all other wind farms in its cluster; normalizing the average feature similarity to obtain the fusion weight of each wind farm.
[0030] Based on the feature similarity S between the wind farm and other stations within the cluster j The average similarity of characteristic parameters between the cluster and other wind farms is selected as the representative index, and the normalized weight W is calculated. j The higher the similarity, the greater the weight. The formula is:
[0031] Where Wj is the weight of the j-th wind farm, and Sj is the weight of the j-th wind farm relative to all other wind farms in the cluster. The average similarity of wind farms, where M is the number of wind farms in the cluster.
[0032] In one specific embodiment, the initial set value of each wind farm cluster is calculated using the following formula:
[0033] in, This represents the initial set of values for the wind farm cluster. Let be the weight of the j-th wind farm in the wind farm cluster; Let be the predicted individual reserve capacity of the j-th wind farm; M is the total number of wind farms in this wind farm cluster. The core of this scheme is to offset the differences in turbine models and human intervention errors in individual models through multi-member fusion.
[0034] Understandably, relying solely on internal consultations may have limitations; for example, the entire cluster's model might be insufficiently prepared for certain rare weather conditions. Therefore, a baseline prediction model for the icing reserve capacity of the wind farm cluster is needed, allowing this model to provide predictions based on long-term data. Finally, by using a fusion correction formula, the results of the internal consultations are combined with historical benchmarks to generate the final prediction value for the cluster. Repeating the above steps for the N divided clusters yields the predicted icing reserve capacity for each cluster. Then, summing the predicted icing reserve capacity for all wind farm clusters provides the total predicted icing reserve capacity for the region.
[0035] The technical solution provided in this application is based on the principle that wind fields are influenced by large-scale atmospheric circulation systems and local topographic elements. If the current wind field is similar to historical wind fields and their dynamic evolution is similar, then the output power of the wind farms will be similar. To eliminate similarity errors, this application proposes using ensemble averaging for forecasting, eliminating the randomness of similarity forecasts and making the forecast results more accurate. Specifically, the technical solution of this application spatially breaks down the large, heterogeneous wind farm cluster into several homogeneous sub-regions through clustering. A two-layer prediction architecture of individual model + benchmark model is also constructed for each sub-region. Within each sub-region, a weighted average is used to integrate individual opinions to smooth out random errors, and then benchmark correction is used to introduce overall patterns to correct systemic biases. This series of operations systematically and hierarchically solves the two major problems pointed out in the background technology: inconsistent standards due to differences in turbine models and strong randomness brought about by human intervention. It achieves a fundamental transformation from unpredictable strong randomness of wind farms to predictable cluster regularity.
[0036] In one embodiment, the predicted icing fallback capacity for each wind farm cluster is calculated using the following formula:
[0037] in, Forecast values of icing fallback capacity for wind farm clusters; The fusion coefficient; This represents the initial set of values for the wind farm cluster. This is the baseline forecast value for the wind farm cluster.
[0038] In this embodiment, to avoid common errors among the aggregate members, such as all individual station models underestimating the icing intensity, leading to C...ensemble Deviation can be introduced into the output C of the cluster baseline model. base The corrections are as shown in the formula above.
[0039] In a specific embodiment, α is the fusion coefficient, which can be determined through optimization using validation set data. A larger value for α ensures that the error cancellation effect of the fusion set is dominant; (1) α) Take a smaller value to correct common deviations in set fusion.
[0040] Repeat the above steps to calculate the predicted ice cover backup capacity C1, C2, ..., C for each of the N clusters in the future time period. N By summarizing the prediction results of all clusters, the predicted value of the regional wind power icing backup capacity is obtained. :
[0041] In one embodiment, the historical aggregated data includes at least: time series aggregation of historical icing monitoring data of all wind farms in the cluster, time series aggregation of historical decommissioning capacity records, and time series of corrected meteorological data covering the area where the cluster is located; the individual icing decommissioning capacity prediction model is trained based on the following data of the corresponding wind farm: historical icing event data of the wind farm, historical decommissioning operation record data, and time series of measured meteorological data at the wind farm site.
[0042] In this embodiment, all wind farms within each cluster can be considered as ensemble prediction members. Each member has an independent single-site icing and reserve capacity prediction model. This model is trained based on the wind farm's own historical icing data, historical reserve capacity, and site-level meteorological data. This single-site model is used to output the individual predicted value C for that wind farm. j Its function is to provide multi-source initial inputs for ensemble prediction, but it does not directly summarize them. Simultaneously, based on cluster aggregated data, including but not limited to historical icing data aggregation of all wind farms, historical reserve capacity aggregation, and cluster-corrected meteorological data, a cluster-level icing reserve capacity baseline prediction model is trained, outputting the cluster reserve capacity baseline value C. base The core function of this benchmark model is to provide a basic anchor point that closely approximates the true value based on the overall pattern of the cluster, thus avoiding deviations in the fusion result caused by excessive individual errors of the ensemble prediction members.
[0043] In summary, the beneficial effects of this application are as follows: 1. Wind farms within the same cluster exhibit highly similar icing environments, and their core driving factors for icing shutdown are consistent. Differences in shutdown thresholds due to turbine model variations can be offset by multi-member fusion of ensemble predictions—the shutdown behavior of different turbine models is considered a reasonable fluctuation within the cluster, and the fused cluster prediction results better reflect the true icing shutdown pattern, improving prediction accuracy. Although the manual shutdown behavior of wind farms within the same cluster has individual randomness, the overall trend of manual intervention is consistent due to similar meteorological and topographical conditions. Ensemble prediction smooths individual random fluctuations through the collective wisdom of multiple wind farms, making the predicted value of cluster shutdown capacity closer to the actual situation and effectively reducing prediction bias caused by manual intervention.
[0044] Figure 1 This is a flowchart illustrating a clustered ensemble prediction method for regional wind power icing back-off reserve capacity in one embodiment. It should be understood that, although... Figure 1 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise explicitly stated herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 1 At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.
[0045] In one embodiment, a regional wind power icing backup capacity clustered prediction device (not shown in the figure) is provided, the device comprising: The memory is configured to store instructions; The processor is configured to retrieve instructions from memory and, when executing the instructions, to implement the regional wind power icing backup capacity clustered prediction method of any of the above embodiments.
[0046] The processor contains a kernel, which retrieves the corresponding program unit from memory. One or more kernels can be configured, and the clustered prediction method for regional wind power icing backup capacity can be implemented by adjusting kernel parameters.
[0047] The memory may include non-permanent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.
[0048] This application provides a storage medium storing a program that, when executed by a processor, implements the aforementioned method for clustered prediction of wind power icing back-up capacity in a given region.
[0049] This application provides a processor for running a program, wherein the program executes the above-mentioned regional wind power icing backup capacity clustered prediction method during runtime.
[0050] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 2 As shown, the computer device includes a processor A01, a network interface A02, a memory (not shown), and a database (not shown) connected via a system bus. The processor A01 provides computing and control capabilities. The memory includes internal memory A03 and a non-volatile storage medium A04. The non-volatile storage medium A04 stores an operating system B01, a computer program B02, and a database (not shown). The internal memory A03 provides an environment for the operation of the operating system B01 and the computer program B02 stored in the non-volatile storage medium A04. The network interface A02 is used for communication with external terminals via a network connection. When the computer program B02 is executed by the processor A01, it implements a clustered prediction method for regional wind power icing backup capacity.
[0051] Those skilled in the art will understand that Figure 2 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.
[0052] This application provides a computer (electronic) device, which includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps of any of the above-mentioned regional wind power icing backup capacity clustered ensemble prediction methods.
[0053] This application also provides a computer program product that, when executed on a data processing device, is suitable for performing the steps of initializing a clustered prediction method for regional wind power icing back-up capacity.
[0054] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0055] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0056] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0057] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0058] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.
[0059] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.
[0060] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.
[0061] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. 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 apparatus that includes that element.
[0062] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.
Claims
1. A clustered prediction method for regional wind power icing backup capacity, characterized in that, The method includes: Based on meteorological and topographical features, wind farms in the region are clustered into multiple wind farm clusters. For each wind farm cluster, an individual icing reserve capacity prediction model is established for each wind farm within the cluster, and a baseline prediction model for the icing reserve capacity of the wind farm cluster is established based on the historical aggregated data of the wind farm cluster. The individual icing-induced reserve capacity prediction model is used to obtain the individual reserve capacity prediction values for each wind farm within each wind farm cluster. For each wind farm cluster, the fusion weight of each wind farm in the wind farm cluster is determined based on the similarity of meteorological and topographic feature parameters of each wind farm in the wind farm cluster. The fusion weight is then used to perform weighted fusion of the individual reserve capacity prediction values of each wind farm to obtain the preliminary set value of the wind farm cluster. The baseline prediction value of each wind farm cluster is obtained by using the baseline prediction model of each wind farm cluster, and the preliminary value of the cluster is fused and corrected with the baseline prediction value to obtain the icing backup capacity prediction value of each wind farm cluster. The total predicted wind power icing reserve capacity of the region is obtained by summing the predicted values of the icing reserve capacity of all wind farm clusters.
2. The method according to claim 1, characterized in that, The method of clustering wind farms within a region into multiple wind farm clusters based on meteorological and topographical features includes: Acquire meteorological and topographic data of each wind farm in the region during a preset historical period; Feature parameters for clustering are selected from the meteorological and topographic data, wherein the meteorological feature parameters include at least the average daily minimum temperature, average daily relative humidity and average daily wind speed in winter, and the topographic feature parameters include at least the average altitude, average slope and landform type quantification value. The selected feature parameters are standardized. A clustering algorithm is used to calculate the Euclidean distance between each wind farm based on the standardized feature parameters, and the wind farms in the region are divided into multiple wind farm clusters according to the Euclidean distance.
3. The method according to claim 1, characterized in that, The fusion weights for each wind farm are determined based on the similarity between the meteorological and topographical features parameters. Calculate the average feature similarity between each wind farm and all other wind farms in its cluster; The average value of the feature similarity is normalized to obtain the fusion weight of each wind farm.
4. The method according to claim 1, characterized in that, The initial set of values for each wind farm cluster is calculated using the following formula: in, This is the initial set value of the wind farm cluster; Let be the weight of the j-th wind farm in the wind farm cluster; Let be the predicted individual reserve capacity of the j-th wind farm; M is the total number of wind farms in the wind farm cluster.
5. The method according to claim 1, characterized in that, The predicted icing fallback capacity for each wind farm cluster is calculated using the following formula: in, The predicted value of the icing back-off capacity of the wind farm cluster; The fusion coefficient; This is the initial set value of the wind farm cluster; For the benchmark pre- Measured value.
6. The method according to claim 5, characterized in that, The fusion coefficient The determination was made through optimization using historical validation set data.
7. The method according to claim 1, characterized in that, The historical aggregated data includes at least: time series aggregation of historical icing monitoring data of all wind farms in the cluster, time series aggregation of historical reserve capacity records, and corrected meteorological data time series covering the area where the cluster is located; The individual icing standby capacity prediction model is trained based on the following data from the corresponding wind farm: Historical icing event data, historical shutdown operation records, and time series of measured meteorological data at the wind farm site.
8. A clustered prediction device for regional wind power icing backfill capacity, characterized in that, include: The memory is configured to store instructions; The processor is configured to retrieve the instructions from the memory and, when executing the instructions, to implement the regional wind power icing backup capacity clustered prediction method according to any one of claims 1 to 7.
9. A machine-readable storage medium storing instructions thereon, characterized in that, When executed by a processor, this instruction causes the processor to be configured to perform the regional wind power icing back-off capacity clustered ensemble prediction method according to any one of claims 1 to 7.
10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the regional wind power icing back-up capacity clustered prediction method according to any one of claims 1 to 7.