Method and system for assessing climate impact of cascade hydropower storage based on meteorological data

Abstract

This invention discloses a method and system for assessing the climate impact of cascade hydropower station storage based on meteorological data, belonging to the field of intelligent management technology. The method includes: performing low-dimensional vectorization processing on historical meteorological data of the target cascade hydropower station using a low-dimensional dense data embedding method; identifying climate impact anchor elements and conducting climate evolution impact analysis; conducting response characteristic assessment experiments based on a set of typical climate scenarios; and combining the results of climate evolution and typical climate impact analysis to comprehensively obtain the climate impact assessment results for the target storage. This invention solves the technical problem that existing climate impact assessment methods rely on static models and extensive calculations, making it difficult to achieve accurate and rapid assessment of the impact of cascade hydropower station storage under variable climate conditions. It achieves the technical effect of improving the assessment efficiency and accuracy of the impact of cascade hydropower station storage under variable climate conditions through low-dimensional data embedding and anchor element identification.

Description

Technical Field

[0001] This invention relates to the field of intelligent management technology, specifically to a method and system for assessing the climate impact of cascade hydropower storage based on meteorological data. Background Technology

[0002] The impacts of climate change on water resource management and hydropower production are becoming increasingly significant. As crucial energy facilities, the water storage and scheduling of cascade hydropower stations directly affect power generation efficiency and regional water resource balance. However, existing climate impact assessment methods often rely on static models and extensive calculations, failing to quickly and accurately reflect the specific impacts of climate change on the water storage of cascade hydropower stations and making it difficult to address changes under complex climate scenarios. Furthermore, these methods are inadequate in dynamically monitoring meteorological data, limiting the real-time optimization of hydropower station operation strategies. Summary of the Invention

[0003] This application provides a method and system for assessing the climate impact of cascade hydropower storage based on meteorological data. This method addresses the technical problem that existing climate impact assessment methods rely on static models and extensive calculations, making it difficult to achieve accurate and rapid assessment of the impact of cascade hydropower storage on variable climate conditions.

[0004] The first aspect of this application provides a method for assessing the climate impact of cascade hydropower storage based on meteorological data. The method includes: using a low-dimensional dense data embedding method to perform low-dimensional floating-point vectorization processing on a set of historical meteorological observation data for a target cascade hydropower station set, obtaining a set of historical low-dimensional floating-point vector sequences; identifying anchor elements based on the set of historical low-dimensional floating-point vector sequences, and performing a union operation on the set of anchor element identification results to determine a set of climate impact anchor elements; using the set of climate impact anchor elements as a guide, performing climate evolution impact analysis on the set of historical low-dimensional floating-point vector sequences, determining a set of climate evolution impact analysis results; traversing the target cascade hydropower station set to conduct response characteristic assessment experiments according to a preset set of typical scenarios, and performing response characteristic assessment based on the response characteristic assessment experiment data set to determine a set of typical climate impact analysis results; and comprehensively analyzing the set of climate evolution impact analysis results and the set of typical climate impact analysis results to obtain the climate impact assessment result for the target hydropower storage.

[0005] The second aspect of this application provides a climate impact assessment system for cascade hydropower storage based on meteorological data. The system includes: a low-dimensional floating-point vectorization processing module, used to perform low-dimensional floating-point vectorization processing on the historical meteorological observation data set of the target cascade hydropower station set using a low-dimensional dense data embedding method to obtain a set of historical low-dimensional floating-point vector sequences; an anchor element identification module, used to identify anchor elements based on the set of historical low-dimensional floating-point vector sequences, and to perform a union operation on the anchor element identification results set to determine the set of climate impact anchor elements; and a climate evolution impact analysis module, used for... Guided by the set of climate impact anchor elements, the set of historical low-dimensional floating-point vector sequences is subjected to climate evolution impact analysis to determine the set of climate evolution impact analysis results. The response feature evaluation module is used to traverse the set of target cascade hydropower stations and conduct response feature evaluation experiments according to a preset set of typical scenarios, and to evaluate the response features based on the set of response feature evaluation experiment data to determine the set of typical climate impact analysis results. The water storage climate impact evaluation module is used to comprehensively analyze the set of climate evolution impact analysis results and the set of typical climate impact analysis results to obtain the target water storage climate impact evaluation results.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0007] This application provides a method and system for assessing the climate impact of cascade hydropower station water storage based on meteorological data, which relates to the field of intelligent management technology. It utilizes a low-dimensional dense data embedding method to perform low-dimensional vectorization processing of meteorological data, guides climate evolution analysis by identifying climate impact anchor elements, assesses response characteristics in conjunction with typical climate scenarios, and comprehensively analyzes the results of climate evolution and typical climate impacts to accurately assess the climate impact of cascade hydropower station water storage. This solves the technical problem that existing climate impact assessment methods rely on static models and extensive calculations, making it difficult to achieve accurate and rapid assessment of the impact of cascade hydropower station water storage under variable climate conditions. It achieves the technical effect of improving the assessment efficiency and accuracy of the impact of cascade hydropower station water storage under variable climate conditions through low-dimensional data embedding and anchor element identification. Attached Figure Description

[0008] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 A schematic diagram of the process for assessing the climate impact of cascade hydropower storage based on meteorological data, provided in the embodiments of this application;

[0010] Figure 2 A schematic diagram of the structure of a cascade hydropower storage climate impact assessment system based on meteorological data, provided in an embodiment of this application.

[0011] Figure labeling: Low-dimensional floating-point number vectorization processing module 11, anchor element identification module 12, climate evolution impact analysis module 13, response characteristic assessment module 14, water storage climate impact assessment module 15. Detailed Implementation

[0012] This application provides a method and system for assessing the climate impact of cascade hydropower storage based on meteorological data. This method addresses the technical problem that existing climate impact assessment methods rely on static models and extensive calculations, making it difficult to achieve accurate and rapid assessment of the impact of cascade hydropower storage on variable climate conditions.

[0013] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. 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.

[0014] It should be noted that the terms "first," "second," etc., in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to such processes, methods, products, or devices.

[0015] Example 1, as Figure 1 As shown, this application provides a method for assessing the climate impact of cascade hydropower storage based on meteorological data, the method comprising:

[0016] P10: Using a low-dimensional dense data embedding method, the historical meteorological observation data set of the target cascade hydropower station set is processed into low-dimensional floating-point vectorization to obtain a set of historical low-dimensional floating-point vector sequences.

[0017] Furthermore, step P10 in this embodiment of the application also includes:

[0018] P11: Using an autoencoder as a framework, collect sample training data to train the framework and obtain a low-dimensional dense data converter; P12: Input the historical meteorological observation data set into the low-dimensional dense data converter in chronological order to perform low-dimensional floating-point vectorization conversion and obtain a set of historical low-dimensional floating-point vector sequences.

[0019] It should be understood that by using low-dimensional dense data embedding methods to process the historical meteorological observation data set of the target cascade hydropower station set, the high-dimensional and complex meteorological data is transformed into a low-dimensional floating-point vector form, thereby achieving data dimensionality reduction and facilitating subsequent analysis and processing. In the specific implementation process, an autoencoder is first used as a framework to train the historical meteorological data. An autoencoder is a neural network model that can compress high-dimensional data into a low-dimensional representation through the encoder part of the input data, and then reconstruct it through the decoder. In this process, the autoencoder learns the implicit patterns and features in the data, removes redundant information from the original meteorological data, and retains the most critical climate features.

[0020] The training data for the autoencoder comes from historical meteorological observation data sets of the target cascade hydropower stations. This data covers multiple meteorological elements such as temperature, precipitation, and wind speed, and is presented in chronological order. The trained autoencoder model can effectively perform low-dimensional mapping on the original data, transforming each set of high-dimensional meteorological data into a low-dimensional, dense floating-point vector. In this way, the structure of the meteorological data is simplified, while key climate information is still preserved.

[0021] Next, a pre-trained low-dimensional dense data transformer is used to perform low-dimensional floating-point vectorization on the historical meteorological observation data. Since meteorological data inherently has a temporal order, the conversion must maintain the chronological sequence. Specifically, the meteorological observation data for each moment is input into a trained autoencoder for conversion, obtaining a low-dimensional floating-point vector corresponding to each time point. Through this process, all historical meteorological data are transformed into a sequence of low-dimensional floating-point vectors. These sequences effectively represent the key information in the historical meteorological data and facilitate subsequent analysis of climate evolution impacts and assessment of response characteristics.

[0022] Finally, the resulting set of low-dimensional floating-point vector sequences is used as input data, providing a foundation for subsequent analysis of climate evolution impacts and supporting further model calculations to assess climate impacts under typical climate scenarios. This processing method not only improves data processing efficiency but also ensures the accuracy and operability of climate change analysis.

[0023] Furthermore, step P10 in this embodiment of the application also includes:

[0024] P13a: Input the historical low-dimensional floating-point vector sequence set into the automatic decoder to obtain the reconstructed historical meteorological observation data set; P14a: Determine the reconstruction error coefficient set based on the reconstruction error between the reconstructed historical meteorological observation data set and the historical meteorological observation data set; P15a: Perform similar aggregation on the historical meteorological observation data set using the historical low-dimensional floating-point vector sequence set, calculate the silhouette coefficient based on the similar aggregation result, and add the difference between 1 and the silhouette coefficient to the clustering error coefficient set; P16a: Perform weighted fusion on the reconstruction error coefficient set and the clustering error coefficient set, and update and optimize the network parameters of the low-dimensional dense data converter based on the fusion result.

[0025] Optionally, the low-dimensional embedding can be further reconstructed and evaluated to ensure that the key features of the original meteorological data are effectively preserved.

[0026] First, a set of historical low-dimensional floating-point vector sequences is input into an automatic decoder for reconstruction, with the aim of recovering approximate historical meteorological observation data from the low-dimensional representation. The decoder, based on pre-trained low-dimensional dense vectors, reconstructs high-dimensional meteorological data as closely as possible to the original data. This process allows for the evaluation of the effectiveness of the low-dimensional data transformation, specifically the difference between the reconstructed data and the original data.

[0027] Next, a set of reconstruction error coefficients is calculated based on the differences between the reconstructed historical meteorological observation dataset and the original dataset. These error coefficients reflect the degree of difference between the meteorological data recovered from the low-dimensional floating-point vector and the original meteorological data. The magnitude of the reconstruction error coefficients helps to understand the information loss or errors during the low-dimensional transformation process and is used for subsequent model optimization.

[0028] Next, a clustering operation is performed on the historical low-dimensional floating-point vector sequence set, aiming to group historical meteorological data with similar climatic characteristics into clusters. Clustering groups similar patterns together, providing valuable structural information for subsequent analysis. The quality of the clustering results is measured by calculating the silhouette coefficient; a higher silhouette coefficient indicates better clustering. Then, the difference between 1 and the silhouette coefficient is added to the clustering error coefficient set to quantify the error in the clustering process.

[0029] Finally, the set of reconstruction error coefficients and the set of clustering error coefficients are weighted and fused. This weighted fusion process comprehensively considers both reconstruction and clustering errors. By assigning weights to these errors and performing weighted calculations, a comprehensive evaluation index is generated to reflect the overall error level during the low-dimensional data transformation process. Based on this fusion result, the network parameters of the low-dimensional dense data converter are updated and optimized. Optimizing the network parameters improves the model's accuracy in data reconstruction and clustering analysis, thereby enhancing the accuracy of subsequent climate evolution analysis and climate impact assessment.

[0030] By following the steps above, the low-dimensional data converter can be effectively optimized, enabling it to better process historical meteorological data, reduce errors, and ensure higher-quality data support in subsequent climate change analysis.

[0031] P20: Based on the set of historical low-dimensional floating-point vector sequences, anchor elements are identified respectively, and the set of anchor element identification results is combined to determine the set of climate influence anchor elements.

[0032] Furthermore, step P20 in this embodiment of the application also includes:

[0033] P21: Analyze the first historical low-dimensional floating-point vector sequence in the set of historical low-dimensional floating-point vector sequences according to the multi-anchor point identification scale set to determine the first historical low-dimensional floating-point asynchronous feature vector set, wherein the first historical low-dimensional floating-point vector sequence is any historical low-dimensional floating-point vector sequence in the set of historical low-dimensional floating-point vector sequences; P22: Perform feature vector interaction enhancement on each pair of historical low-dimensional floating-point asynchronous feature vectors in the first historical low-dimensional floating-point asynchronous feature vector set, and perform mean processing on the enhanced historical low-dimensional floating-point asynchronous enhanced feature vectors to determine the first historical low-dimensional floating-point asynchronous enhanced feature vector; P23: Add the low-dimensional floating-point elements corresponding to the first m low-dimensional floating-point asynchronous enhanced feature elements in the first historical low-dimensional floating-point asynchronous enhanced feature vector to the climate influence anchor point element set.

[0034] Specifically, anchor elements are identified based on the processed set of historical low-dimensional floating-point vector sequences. Anchor elements here refer to key data points that significantly represent climate change characteristics. By analyzing the set of historical low-dimensional floating-point vector sequences, these anchor points—elements that effectively capture important characteristics of climate change—are identified. All identified anchor elements are then combined using a union operation to ultimately determine a set of climate impact anchor elements.

[0035] First, for each individual sequence in the set of historical low-dimensional floating-point vector sequences (e.g., the first historical low-dimensional floating-point vector sequence), a multi-anchor recognition scale set is applied for analysis. The core of this process is to analyze the differences in the changes of each element within each vector at different recognition scales, aiming to identify significant features that change over time. Multi-anchor recognition scales refer to observing data through different scales or perspectives (possibly time windows or other scales), that is, analyzing the degree of difference in the changes of each element within the vector over time according to the multi-anchor recognition scale set, thereby capturing the changing trends within different periods. Through this method, the first historical low-dimensional floating-point asynchronous feature vector set is finally obtained, which reflects the differences in the changes of each element in the vector over time.

[0036] Next, to further improve the analysis results, feature vector interaction enhancement was performed on each pair of vectors in the first historical low-dimensional floating-point asynchronous feature vector set. This involved combining asynchronous feature vectors from different time points to uncover potential correlations and common features among them. The enhanced asynchronous feature vectors were then averaged to ensure consistency and representativeness. The averaged feature vector yielded the first historical low-dimensional floating-point asynchronous enhanced feature vector, representing the key features extracted during the interaction enhancement process, which is crucial for subsequent anchor element identification.

[0037] Finally, low-dimensional floating-point asynchronous enhancement feature elements are extracted from the first historical low-dimensional floating-point asynchronous enhancement feature vector. Specifically, the first m low-dimensional floating-point elements in these feature vectors are selected as candidate elements for climate impact anchors. These first m elements are of high importance, representing the most significant features in climate change. By adding these feature elements to the set of climate impact anchor elements, a complete and highly representative set of climate impact anchor elements is finally formed.

[0038] By implementing the above steps, key anchor elements closely related to climate change can be effectively identified, and through multi-scale analysis and interactive enhancement, the identified anchor elements can be ensured to have high credibility and representativeness.

[0039] Furthermore, step P21 in this embodiment of the application also includes:

[0040] P21-1: Perform curve fitting on the set of historical low-dimensional floating-point vector sequences respectively, and extract the extreme points in the fitted curves to obtain a set of extreme point sequences of fitted curves; P21-2: Traverse and count the time intervals between two adjacent extreme points in the set of extreme point sequences of fitted curves to obtain a set of time interval sequences; P21-3: Extract the maximum and minimum time intervals in each time interval sequence in the set of time interval sequences, and perform a union of the extraction results to obtain the set of multi-anchor point recognition scales.

[0041] In one possible embodiment of this application, the process of conducting in-depth analysis of the set of historical low-dimensional floating-point vector sequences can be further refined to determine the set of multi-anchor point identification scales.

[0042] First, curve fitting is performed on each historical low-dimensional floating-point vector sequence. The purpose of curve fitting is to find a smooth curve that mathematically approximates the changing trends of historical meteorological data as closely as possible. This process constructs a fitted curve for each historical vector sequence by applying an appropriate fitting algorithm (such as polynomial fitting or spline fitting). Subsequently, extreme points, i.e., local maxima and minima, are extracted from the fitted curves. These extreme points represent key nodes of change in historical meteorological data and are of great significance for identifying the impacts of climate change. Finally, these extreme points are collected into a set of fitted curve extreme point sequences, which will serve as the basis for subsequent analysis.

[0043] Subsequently, the time intervals between every two adjacent extreme points in the set of extreme point sequences of the fitted curve are traversed and statistically analyzed. The time interval refers to the time difference between adjacent extreme points; by statistically analyzing these time intervals, the rhythm and periodicity of meteorological data changes can be revealed. For example, some meteorological changes may exhibit regular fluctuations, while others may be sudden; by statistically analyzing the time intervals, the regularity of these fluctuations can be better understood. This process produces a set of time interval sequences, where each element represents the time interval between two adjacent extreme points of the fitted curve.

[0044] Finally, the maximum and minimum time intervals in each time interval sequence were extracted from the set of time interval sequences. The maximum and minimum time intervals represent the longest and shortest periods occurring during meteorological changes, respectively. These time intervals reflect the periodic characteristics and fluctuation intensity of climate change. By taking the union of the maximum and minimum time intervals in each time interval sequence, a multi-anchor identification scale set was obtained. This set contains the change scales of all time intervals, representing the main characteristics of climate change at different time scales.

[0045] This coherent analytical process identifies key meteorological features that play a crucial role in climate impacts across different timescales. These features are used as anchor points to determine the set of anchor elements for climate impacts. This not only improves the accuracy of climate impact assessments but also enhances the reliability of the results.

[0046] Furthermore, step P22 in the embodiments of this application also includes:

[0047] P22-1: Randomly extract two first historical low-dimensional floating-point asynchronous feature vectors from the first historical low-dimensional floating-point asynchronous feature vector set; P22-2: Perform vector element similarity parsing on the two first historical low-dimensional floating-point asynchronous feature vectors to determine the vector element similarity set, and perform matrix processing on the vector element similarity set to construct an interaction enhancement guidance matrix; P22-3: Use the interaction enhancement guidance matrix to convolve the two first historical low-dimensional floating-point asynchronous feature vectors respectively to obtain the two enhanced historical low-dimensional floating-point asynchronous feature vectors.

[0048] Optionally, the process of enhancing the feature representation capability of meteorological data through interactive enhancement of feature vectors can be further refined.

[0049] First, two historical low-dimensional floating-point asynchronous feature vectors are randomly extracted from the first set of historical low-dimensional floating-point asynchronous feature vectors. These feature vectors represent the changing characteristics of meteorological data at different time points or time periods. By selecting them randomly, the diversity of the samples can be ensured, which helps to capture the changing patterns in different time series and provides rich input for subsequent feature enhancement and optimization.

[0050] Next, vector element similarity analysis is performed on these two historical low-dimensional floating-point asynchronous feature vectors. By evaluating the degree of similarity between the two vectors at each element, their similarity in meteorological changes is reflected. For example, to quantify this similarity, common similarity calculation methods such as cosine similarity and Euclidean distance can be used. These calculation results will form a vector element similarity set, which contains the similarity information of each element between the two feature vectors.

[0051] Finally, the two extracted historical low-dimensional floating-point asynchronous feature vectors are convolved using the constructed interactive enhancement guidance matrix. Convolution is a common signal processing technique that extracts important features within local regions of a vector using a sliding window approach. Here, through convolution, the two historical low-dimensional floating-point asynchronous feature vectors are combined with the interactive enhancement guidance matrix to further enhance their performance on similar elements, resulting in enhanced historical low-dimensional floating-point asynchronous feature vectors. These enhanced feature vectors not only retain the original meteorological change characteristics but also improve their representativeness through the interactive enhancement process, making them more capable of describing climate change.

[0052] Through the above steps, the feature vectors are not only enhanced, but the relationships between features are also strengthened through similarity parsing and convolution. This provides more accurate and representative data support for subsequent climate change analysis, model training and evaluation.

[0053] P30: Guided by the set of climate impact anchor elements, conduct climate evolution impact analysis on the set of historical low-dimensional floating-point vector sequences respectively, and determine the set of climate evolution impact analysis results.

[0054] Furthermore, step P30 in this embodiment of the application also includes:

[0055] P31: Guided by the set of climate impact anchor elements, conduct a guided correlation analysis on the set of historical low-dimensional floating-point vector sequences to construct a set of historical low-dimensional floating-point vector guided correlation matrix sequences; P32: Use long and short time series analysis networks to conduct long and short time series iterative analysis on the set of historical low-dimensional floating-point vector guided correlation matrix sequences to determine the set of climate evolution impact analysis results.

[0056] Specifically, the set of climate impact anchor elements is used to guide the analysis of the climate evolution impact of historical low-dimensional floating-point vector sequences, thereby determining the set of climate evolution impact analysis results.

[0057] First, a correlation analysis is conducted on a set of historical low-dimensional floating-point vector sequences using a set of climate impact anchor elements as a guide. This set of anchor elements contains key climate change characteristics; therefore, using these elements to guide correlation analysis can uncover potential patterns related to climate change from meteorological data. In this step, a set of historical low-dimensional floating-point vector-guided correlation matrix sequences is constructed by calculating the correlations between the historical floating-point vectors. This matrix sequence reflects the interrelationships between different time points in historical meteorological data and their connection with climate change, reflecting the degree of impact of climate change on meteorological data.

[0058] Next, long and short time series analysis networks (such as LSTM networks) are used to perform iterative long and short time series analysis on the historical low-dimensional floating-point vector-guided correlation matrix sequence set. Long and short time series analysis networks are deep learning models specifically designed for processing and predicting time series data, capable of capturing long-term dependencies in long-term series. Through LSTM networks, long-term climate evolution trends can be effectively extracted from historical meteorological data and subjected to time series analysis. The LSTM model iteratively learns and optimizes the relationship between historical data and climate change, thus yielding more accurate results in analyzing the impacts of climate change. Ultimately, this analysis generates a set of climate change impact analysis results, reflecting the impact of climate change on various historical time points and providing data support for further climate change assessment and prediction.

[0059] P40: Traverse the target cascade hydropower station set and conduct response characteristic evaluation tests according to the preset typical scenario set, and conduct response characteristic evaluation based on the response characteristic evaluation test data set to determine the typical climate impact analysis result set.

[0060] Furthermore, step P40 in this embodiment of the application also includes:

[0061] P41: Obtain a preset set of typical scene types, and perform scene data mining based on the preset set of typical scene types to obtain a preset set of typical scene mining data; P42: Based on the preset set of typical scene types, perform mean shift analysis on the preset set of typical scene mining data for each preset typical scene type to determine the preset set of typical scenes.

[0062] Optionally, a response characteristic assessment of climate change can be conducted based on the target cascade hydropower station set, thereby generating a set of typical climate impact analysis results.

[0063] First, a set of pre-defined typical scenario types is obtained. These scenarios represent various situations that may affect the operation of hydropower stations and climate change, such as extreme weather, drought periods, or periods of heavy rainfall. Through scenario data mining, these typical scenarios are analyzed in depth to uncover key meteorological data and potential climate patterns related to each scenario type. The pre-defined typical scenario mining dataset obtained through scenario data mining contains specific meteorological variable data for all these scenarios, facilitating subsequent analysis and evaluation.

[0064] Next, based on the aforementioned set of preset typical scenario types, the dataset of preset typical scenario mining is further processed. Specifically, for each preset typical scenario type, mean shift analysis is performed. Mean shift analysis is a non-parametric statistical method typically used to find high-density regions in data. By analyzing the data for each scenario type, it identifies the common data characteristics of that scenario. In this step, mean shift analysis can be used to determine the common data corresponding to each typical scenario type, that is, the most common or representative data characteristics in that scenario.

[0065] Mean shift analysis can identify the center or mean position of each pre-defined typical scenario type, representing the typical data distribution under that scenario. This analysis determines the general data characteristics corresponding to each typical scenario type, thus forming a pre-defined set of typical scenarios encompassing all typical scenarios. This set will provide a solid foundation of data for subsequent climate impact analysis and hydropower station response characteristic assessment.

[0066] Finally, response characteristic assessment experiments were conducted on the target cascade hydropower station set using a pre-set set of typical scenarios. By evaluating the response characteristics of the hydropower stations to different climate scenarios, including their adaptability, vulnerability, and potential impacts, a response characteristic assessment experiment dataset was obtained. Based on this dataset, a response characteristic assessment was then performed to determine the set of typical climate impact analysis results. This dataset provides detailed information on the response characteristics of the cascade hydropower stations under different climate conditions, which helps in developing corresponding management and adaptation strategies.

[0067] P50: By comprehensively analyzing the set of results of the climate evolution impact analysis and the set of results of the typical climate impact analysis, the climate impact assessment results of the target water storage are obtained.

[0068] Specifically, a comprehensive analysis of the set of results for climate evolution impact analysis and the set of results for typical climate impact analysis is conducted to obtain the climate impact assessment results for the target water storage.

[0069] First, the set of results for climate evolution impact analysis and the set of results for typical climate impact analysis represent the evolutionary impacts of climate change on meteorological elements and the impact analysis results under specific typical climate scenarios, respectively. The set of results for climate evolution impact analysis reflects the gradual trend of climate change over a long period and its impact on meteorological data, while the set of results for typical climate impact analysis analyzes the response to climate change based on different typical climate scenarios (such as drought, rainstorms, and warming), revealing the response characteristics of hydropower stations under various climatic conditions.

[0070] Next, these two sets of results will be comprehensively analyzed. First, the results sets from the climate evolution impact analysis and typical climate impact analysis will undergo data preprocessing and standardization to ensure that both sets of data have a consistent temporal, spatial, and contextual basis. During this process, noise and outliers will be removed from the data, and the data will be normalized to ensure that all data items have the same dimensions, thus facilitating subsequent comparison and analysis. Then, weights will be assigned to the two sets of data according to the degree of impact of different climate scenarios. For example, the climate evolution analysis results may reflect long-term trends, while the typical climate analysis results focus on short-term extreme climate impacts; therefore, different weights will be considered for long-term and short-term impacts when assigning weights.

[0071] Based on the weighted data fusion, cluster analysis and pattern recognition methods were used to further identify the main impact patterns of climate evolution and typical climate effects on hydropower stations. K-means clustering technology can be used to divide the data into several climate impact patterns, such as drought patterns and extreme precipitation patterns, thereby clarifying the different impact characteristics of climate change on hydropower stations. Principal component analysis (PCA) can be used to extract the most representative climate factors, helping to identify the climate factors that have the most significant impact on hydropower station water storage.

[0072] Finally, by integrating these models, a comprehensive assessment of the climate impact of the target water storage was obtained. This assessment comprehensively reflects the overall impact of climate evolution trends and typical climate events on the water storage and operation of hydropower stations, providing a scientific basis for hydropower station scheduling and climate adaptation management.

[0073] Furthermore, step P50 in this embodiment of the application also includes: P50a: setting a feedback monitoring cycle based on the target water storage climate impact assessment results, and collecting water storage climate impact data on the target cascade hydropower station set according to the feedback monitoring cycle.

[0074] In one possible embodiment of this application, in order to achieve continuous monitoring and assessment of the climate impact of the target cascade hydropower station's collective water storage, a feedback monitoring cycle can be set based on the assessment results of the target water storage climate impact, and water storage climate impact data can be collected on this basis.

[0075] First, based on the climate impact assessment results of the target reservoir, a feedback monitoring cycle is determined. This assessment provides the response characteristics of the hydropower station under different climate scenarios and the extent to which climate change affects its operation. Based on these assessment results, it is possible to reasonably predict future climate change trends that the hydropower station may encounter. By analyzing these trends, a reasonable feedback monitoring cycle is established to ensure that data collection and monitoring can reflect the real-time impact of climate change on the hydropower station.

[0076] The feedback monitoring cycle refers to the monitoring time interval set based on the results of climate impact assessments, which may be adjusted according to different climate change characteristics. The monitoring cycle can be daily, monthly, quarterly, or annual, to ensure sufficient time for data collection and analysis, especially for factors with slow or cyclical climate change.

[0077] After determining the feedback monitoring cycle, the data acquisition phase begins. According to the established monitoring cycle, data on the climate impact of water storage at the target cascade hydropower stations are collected. The collected data typically includes real-time meteorological data, water storage data, reservoir operational status, and other indicators related to climate change. By collecting this data periodically or in real-time, the climate change impacts of the hydropower stations during water storage can be continuously tracked, potential risks from climate change can be identified in a timely manner, and data support can be provided for future climate impact assessments. This process not only ensures the dynamic monitoring of the hydropower stations' climate change response but also provides real-time data for subsequent climate adaptation adjustments and optimization schemes.

[0078] In summary, the embodiments of this application have at least the following technical effects:

[0079] This application employs a low-dimensional dense data embedding method and climate impact anchor element identification to accurately assess the impact of climate change on the water storage and operation of cascade hydropower stations, thereby improving the accuracy and efficiency of climate impact assessment. By combining climate evolution analysis and response characteristic assessment of typical climate scenarios, it achieves dynamic tracking and real-time assessment of climate change, effectively responding to extreme climate events. Through comprehensive analysis of climate evolution and typical climate impact results, it provides a scientific basis for the operation and scheduling of hydropower stations and water resource allocation, optimizing the scheduling efficiency and stability of hydropower stations. At the same time, it provides scientific decision support for hydropower stations to cope with extreme climate events, ensuring their sustainable operation and climate adaptability.

[0080] The technology has achieved the goal of improving the efficiency and accuracy of assessing the impact of cascade hydropower station water storage under variable climate conditions through low-dimensional data embedding and anchor element identification.

[0081] Example 2 is based on the same inventive concept as the meteorological data-based cascade hydropower storage climate impact assessment method in the aforementioned examples, such as... Figure 2 As shown, this application provides a cascade hydropower storage climate impact assessment system based on meteorological data. The system and method embodiments in this application are based on the same inventive concept. The system includes:

[0082] The low-dimensional floating-point vectorization processing module 11 is used to perform low-dimensional floating-point vectorization processing on the historical meteorological observation data set of the target cascade hydropower station set using a low-dimensional dense data embedding method, so as to obtain a set of historical low-dimensional floating-point vector sequences.

[0083] The anchor element identification module 12 is used to identify anchor elements based on the set of historical low-dimensional floating-point vector sequences, and to perform a union of the anchor element identification result sets to determine the set of climate-affected anchor elements.

[0084] The climate evolution impact analysis module 13 is used to perform climate evolution impact analysis on the historical low-dimensional floating-point vector sequence set, guided by the set of climate impact anchor elements, and to determine the set of climate evolution impact analysis results.

[0085] The response characteristic assessment module 14 is used to traverse the target cascade hydropower station set to conduct response characteristic assessment tests according to a preset typical scenario set, and to conduct response characteristic assessment based on the response characteristic assessment test data set to determine the typical climate impact analysis result set.

[0086] The water storage climate impact assessment module 15 is used to comprehensively analyze the set of climate evolution impact analysis results and the set of typical climate impact analysis results to obtain the target water storage climate impact assessment results.

[0087] Furthermore, the low-dimensional floating-point vectorization processing module 11 is also used to perform the following steps:

[0088] Using an autoencoder as a framework, sample training data is collected to train the framework and obtain a low-dimensional dense data converter. The historical meteorological observation data set is then input into the low-dimensional dense data converter in chronological order to perform low-dimensional floating-point vectorization conversion, resulting in a set of historical low-dimensional floating-point vector sequences.

[0089] Furthermore, the low-dimensional floating-point vectorization processing module 11 is also used to perform the following steps:

[0090] The historical low-dimensional floating-point vector sequence set is input into an automatic decoder to obtain a reconstructed historical meteorological observation data set. Based on the reconstruction error between the reconstructed historical meteorological observation data set and the historical meteorological observation data set, a reconstruction error coefficient set is determined. The historical low-dimensional floating-point vector sequence set is then used to perform similar aggregation on the historical meteorological observation data set, and the silhouette coefficient is calculated based on the aggregation results. The difference between 1 and the silhouette coefficient is added to the clustering error coefficient set. The reconstruction error coefficient set and the clustering error coefficient set are then weighted and fused, and the network parameters of the low-dimensional dense data converter are updated and optimized based on the fusion result.

[0091] Furthermore, the anchor element identification module 12 is also used to perform the following steps:

[0092] The first historical low-dimensional floating-point vector sequence in the set of historical low-dimensional floating-point vector sequences is parsed according to the multi-anchor point identification scale set to determine the first historical low-dimensional floating-point asynchronous feature vector set, wherein the first historical low-dimensional floating-point vector sequence is any historical low-dimensional floating-point vector sequence in the set of historical low-dimensional floating-point vector sequences; the pairwise historical low-dimensional floating-point asynchronous feature vectors in the first historical low-dimensional floating-point asynchronous feature vector set are enhanced by feature vector interaction, and the enhanced historical low-dimensional floating-point asynchronous enhanced feature vectors are averaged to determine the first historical low-dimensional floating-point asynchronous enhanced feature vector; the low-dimensional floating-point asynchronous enhanced feature elements in the first historical low-dimensional floating-point asynchronous enhanced feature vector that are in the first m positions are added to the climate influence anchor point element set.

[0093] Furthermore, the anchor element identification module 12 is also used to perform the following steps:

[0094] Two first historical low-dimensional floating-point asynchronous feature vectors are randomly extracted from the first historical low-dimensional floating-point asynchronous feature vector set; the vector element similarity of the two first historical low-dimensional floating-point asynchronous feature vectors is analyzed to determine the vector element similarity set, and the vector element similarity set is matrixed to construct an interaction enhancement guidance matrix; the interaction enhancement guidance matrix is ​​used to convolve the two first historical low-dimensional floating-point asynchronous feature vectors respectively to obtain the two enhanced historical low-dimensional floating-point asynchronous feature vectors.

[0095] Furthermore, the anchor element identification module 12 is also used to perform the following steps:

[0096] Curve fitting is performed on the set of historical low-dimensional floating-point vector sequences, and the extreme points in the fitted curves are extracted to obtain a set of extreme point sequences of fitted curves. The time intervals between two adjacent extreme points in the set of extreme point sequences of fitted curves are traversed and counted to obtain a set of time interval sequences. The maximum and minimum time intervals in each time interval sequence in the set of time interval sequences are extracted, and the union of the extraction results is obtained to obtain the set of multi-anchor point recognition scales.

[0097] Furthermore, the climate evolution impact analysis module 13 is also used to perform the following steps:

[0098] Guided by the set of climate impact anchor elements, a correlation analysis is performed on the set of historical low-dimensional floating-point vector sequences to construct a set of historical low-dimensional floating-point vector-guided correlation matrix sequences. Long and short time series analysis networks are used to perform long and short time series iterative analysis on the set of historical low-dimensional floating-point vector-guided correlation matrix sequences to determine the set of climate evolution impact analysis results.

[0099] Furthermore, the response feature evaluation module 14 is also used to perform the following steps:

[0100] Obtain a preset set of typical scenario types, perform scenario data mining based on the preset set of typical scenario types to obtain a preset set of typical scenario mining data; based on the preset set of typical scenario types, perform mean shift analysis on the data for each preset typical scenario type in the preset set of typical scenario mining data to determine the preset set of typical scenarios.

[0101] Furthermore, the water storage climate impact assessment module 15 is also used to perform the following steps:

[0102] Based on the target water storage climate impact assessment results, a feedback monitoring cycle is set, and water storage climate impact data are collected for the target cascade hydropower station group according to the feedback monitoring cycle.

[0103] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0104] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0105] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application intends to include such modifications and variations.

Claims

1. A method for assessing the climate impact of cascade hydropower storage based on meteorological data, characterized in that, The method includes: Using a low-dimensional dense data embedding method, the historical meteorological observation data set of the target cascade hydropower station set is processed into low-dimensional floating-point vectorization to obtain a set of historical low-dimensional floating-point vector sequences. This step includes: Using an autoencoder as a framework, sample training data is collected to train the framework, resulting in a low-dimensional dense data converter. The historical meteorological observation data set is input into the low-dimensional dense data converter in chronological order to perform low-dimensional floating-point vectorization conversion, thereby obtaining a set of historical low-dimensional floating-point vector sequences. The sets of historical low-dimensional floating-point vector sequences are input into an automatic decoder to obtain a reconstructed set of historical meteorological observation data. Based on the reconstruction error between the reconstructed historical meteorological observation data set and the historical meteorological observation data set, a set of reconstruction error coefficients is determined. The historical low-dimensional floating-point vector sequence set is aggregated with the historical meteorological observation data set according to the same category, and the silhouette coefficient is calculated based on the aggregation result. The difference between 1 and the silhouette coefficient is added to the clustering error coefficient set. The reconstruction error coefficient set and the clustering error coefficient set are weighted and fused, and the network parameters of the low-dimensional dense data converter are updated and optimized based on the fusion result. Anchor element identification is performed based on the set of historical low-dimensional floating-point vector sequences, and the set of anchor element identification results is combined to determine the set of climate influence anchor elements. Guided by the set of climate impact anchor elements, climate evolution impact analysis is performed on the set of historical low-dimensional floating-point vector sequences to determine the set of climate evolution impact analysis results. The target cascade hydropower station set is traversed to conduct response characteristic evaluation tests according to a preset set of typical scenarios, and response characteristic evaluation is performed based on the set of response characteristic evaluation test data to determine the set of typical climate impact analysis results. By comprehensively analyzing the set of climate evolution impact analysis results and the set of typical climate impact analysis results, the climate impact assessment results of the target water storage are obtained.

2. The method for assessing the climate impact of cascade hydropower storage based on meteorological data as described in claim 1, characterized in that, Anchor element identification is performed based on the historical low-dimensional floating-point vector sequence set, and the union of the anchor element identification result sets is obtained to determine the set of climate influence anchor elements, including: The first historical low-dimensional floating-point vector sequence in the set of historical low-dimensional floating-point vector sequences is analyzed according to the multi-anchor point identification scale set to determine the first historical low-dimensional floating-point asynchronous feature vector set, wherein the first historical low-dimensional floating-point vector sequence is any historical low-dimensional floating-point vector sequence in the set of historical low-dimensional floating-point vector sequences. The first set of historical low-dimensional floating-point asynchronous feature vectors is subjected to feature vector interaction enhancement between each pair of historical low-dimensional floating-point asynchronous feature vectors, and the enhanced historical low-dimensional floating-point asynchronous feature vectors are averaged to determine the first historical low-dimensional floating-point asynchronous feature vector. The low-dimensional floating-point asynchronous enhancement feature elements that are in the first m positions of the first historical low-dimensional floating-point asynchronous enhancement feature vector are added to the climate influence anchor element set.

3. The method for assessing the climate impact of cascade hydropower storage based on meteorological data as described in claim 2, characterized in that, include: Two asynchronous feature vectors of the first historical low-dimensional floating-point numbers are randomly extracted from the first set of historical low-dimensional floating-point feature vectors. The vector element similarity is parsed for two first historical low-dimensional floating-point asynchronous feature vectors to determine the vector element similarity set, and the vector element similarity set is matrixed to construct an interaction enhancement guidance matrix; The two first historical low-dimensional floating-point asynchronous feature vectors are convolved using the interaction enhancement guidance matrix to obtain the two enhanced historical low-dimensional floating-point asynchronous feature vectors.

4. The method for assessing the climate impact of cascade hydropower storage based on meteorological data as described in claim 2, characterized in that, The first historical low-dimensional floating-point vector sequence in the set of historical low-dimensional floating-point vector sequences is analyzed according to the multi-anchor point identification scale set to determine the first historical low-dimensional floating-point asynchronous feature vector set, including: Curve fitting is performed on the set of historical low-dimensional floating-point vector sequences respectively, and the extreme points in the fitted curves are extracted to obtain the set of extreme point sequences of the fitted curves. By iterating through and statistically analyzing the time intervals between two adjacent extreme points of the fitted curve in the set of extreme point sequences of the fitted curve, a set of time interval sequences is obtained. Extract the maximum and minimum time intervals from each time interval sequence in the set of time interval sequences, and perform a union operation on the extraction results to obtain the set of multi-anchor point recognition scales.

5. The method for assessing the climate impact of cascade hydropower storage based on meteorological data as described in claim 1, characterized in that, Guided by the set of climate impact anchor elements, climate evolution impact analysis is performed on the set of historical low-dimensional floating-point vector sequences to determine the set of climate evolution impact analysis results, including: Guided by the set of climate impact anchor elements, a correlation analysis is performed on the set of historical low-dimensional floating-point vector sequences to construct a set of historical low-dimensional floating-point vector guided correlation matrix sequences. Long and short time series analysis networks are used to perform long and short time series iterative analysis on the historical low-dimensional floating-point vector-guided correlation matrix sequence set to determine the set of climate evolution impact analysis results.

6. The method for assessing the climate impact of cascade hydropower storage based on meteorological data as described in claim 1, characterized in that, The target cascade hydropower station set is traversed to conduct response characteristic assessment experiments according to a preset set of typical scenarios. Based on the response characteristic assessment experiment data set, response characteristic assessment is performed to determine the set of typical climate impact analysis results, including: Obtain a preset set of typical scenario types, and perform scenario data mining based on the preset set of typical scenario types to obtain a preset set of typical scenario mining data. Based on the preset set of typical scenario types, mean shift analysis of data for each preset typical scenario type is performed on the preset typical scenario mining data set to determine the preset set of typical scenarios.

7. The method for assessing the climate impact of cascade hydropower storage based on meteorological data as described in claim 1, characterized in that, Based on the target water storage climate impact assessment results, a feedback monitoring cycle is set, and water storage climate impact data are collected for the target cascade hydropower station group according to the feedback monitoring cycle.

8. A cascade hydropower storage climate impact assessment system based on meteorological data, applied to the cascade hydropower storage climate impact assessment method based on meteorological data as described in any one of claims 1-7, characterized in that, The system includes: The low-dimensional floating-point vectorization processing module is used to perform low-dimensional floating-point vectorization processing on the historical meteorological observation data set of the target cascade hydropower station set using a low-dimensional dense data embedding method, and obtain a set of historical low-dimensional floating-point vector sequences. Anchor element identification module is used to identify anchor elements based on the set of historical low-dimensional floating-point vector sequences, and to perform union of the anchor element identification result set to determine the set of climate influence anchor elements. The climate evolution impact analysis module is used to perform climate evolution impact analysis on the set of historical low-dimensional floating-point vector sequences, guided by the set of climate impact anchor points, and to determine the set of climate evolution impact analysis results. The response characteristic assessment module is used to traverse the target cascade hydropower station set to conduct response characteristic assessment tests according to a preset set of typical scenarios, and to conduct response characteristic assessment based on the response characteristic assessment test data set to determine the set of typical climate impact analysis results. The water storage climate impact assessment module is used to comprehensively analyze the set of climate evolution impact analysis results and the set of typical climate impact analysis results to obtain the target water storage climate impact assessment results.

Images