Meteorological data optimization method and device suitable for generalized terrain

Abstract

The invention relates to the technical field of meteorological data processing, and discloses a meteorological data optimization method and device suitable for generalized terrains, and the method comprises the steps: obtaining an initial layout strategy of a wind and light station observation station, carrying out the dynamic site selection optimization of the initial layout strategy of the wind and light station observation station, and obtaining an optimal layout strategy of the wind and light station observation station; obtaining multi-source meteorological data of the target wind and light station based on the optimal layout strategy of the wind and light station observation station, and performing data cleaning on the multi-source meteorological data of the target wind and light station to obtain the multi-source meteorological data after data cleaning; and performing sample expansion on the multi-source meteorological data after data cleaning to obtain a virtual meteorological observation data set. According to the invention, the accuracy of the acquired meteorological data and the meteorological prediction precision of the wind and light station are significantly improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of meteorological data processing, and particularly relates to a meteorological data optimization method and device suitable for generalized terrain. BACKGROUND

[0002] Wind power and photovoltaic power are a kind of volatile and intermittent energy, and large-scale installation and grid connection will inevitably affect the stable operation of the local power grid. During the operation and maintenance of the wind and light field station, the accuracy and reliability of meteorological data are crucial for predicting power generation, formulating maintenance plans and ensuring system safety.

[0003] However, due to the problems such as missing values and abnormal values in the original observed meteorological data, the prediction accuracy and reliability of the meteorological data are affected, and thus the operation and maintenance decision and power generation prediction of the field station are affected. SUMMARY

[0004] Therefore, the present application provides a meteorological data optimization method and device suitable for generalized terrain to solve the problem that the original observed meteorological data is inaccurate and affects the operation and maintenance decision and power generation prediction of the field station.

[0005] In a first aspect, the present application provides a meteorological data optimization method suitable for generalized terrain, which comprises:

[0006] obtaining an initial layout strategy of a wind and light field station, dynamically optimizing the initial layout strategy of the wind and light field station to obtain an optimal layout strategy of the wind and light field station;

[0007] obtaining multi-source meteorological data of a target wind and light field station based on the optimal layout strategy of the wind and light field station, and performing data cleaning on the multi-source meteorological data of the target wind and light field station to obtain cleaned multi-source meteorological data;

[0008] performing sample expansion on the cleaned multi-source meteorological data to obtain a virtual meteorological observation data set.

[0009] The meteorological data optimization method suitable for generalized terrain provided in this embodiment realizes dynamic optimization of the layout of the self-built station by dynamically optimizing the initial layout strategy of the wind and light field station, significantly improves the accuracy of the obtained meteorological data, and improves the operation efficiency of the wind and light field station. Secondly, by performing data cleaning on the multi-source meteorological data of the target wind and light field station, the accuracy and consistency of the data are ensured, and reliable data support is provided for model training and prediction. Finally, by performing sample expansion on the cleaned multi-source meteorological data, a high-resolution virtual meteorological observation data set is formed, the feature expression and real-time monitoring capability of the data are enhanced, and the meteorological prediction accuracy of the wind and light field station is improved.

[0010] In an optional implementation, the initial layout strategy of the wind and light station is dynamically optimized to obtain an optimal layout strategy of the wind and light station, including:

[0011] Based on the initial layout strategy of the wind and light station, environmental parameters and economic configuration parameters of the wind and light station are obtained, and a comprehensive evaluation model is constructed based on the environmental parameters and economic configuration parameters of the wind and light station;

[0012] The initial layout strategy of the wind and light station is evaluated by using the comprehensive evaluation model, and the optimal layout strategy of the wind and light station is determined based on the evaluation result.

[0013] The meteorological data optimization method suitable for generalized terrain provided in this embodiment can efficiently utilize limited observation samples through dynamic site selection optimization, realize economic optimal configuration of station layout, improve the operation efficiency of the station, and reduce the construction and maintenance cost.

[0014] In an optional implementation, the initial layout strategy of the wind and light station is dynamically optimized to obtain an optimal layout strategy of the wind and light station, including:

[0015] Based on the comprehensive evaluation model, a target function of conditional nonlinear optimal disturbance is determined;

[0016] The target function is numerically solved to obtain the optimal layout strategy of the wind and light station.

[0017] The meteorological data optimization method suitable for generalized terrain provided in this embodiment can efficiently utilize limited observation samples through dynamic site selection optimization, realize economic optimal configuration of station layout, improve the operation efficiency of the station, and reduce the construction and maintenance cost.

[0018] In an optional implementation, based on the optimal layout strategy of the wind and light station, multi-source meteorological data of a target wind and light station is obtained, and the multi-source meteorological data of the target wind and light station is data cleaned to obtain data cleaned multi-source meteorological data, including:

[0019] Based on the optimal layout strategy of the wind and light station, a target wind and light station is selected, and multi-source meteorological data of the target wind and light station is obtained;

[0020] The multi-source meteorological data of the target wind and light station is preprocessed to obtain preprocessed meteorological data;

[0021] The multi-source meteorological data of the target wind and light station is preprocessed to obtain preprocessed meteorological data;

[0022] The abnormal meteorological data is corrected to obtain the multi-source meteorological data after data cleaning.

[0023] The method provided by the embodiment is suitable for general topography, and the missing values, abnormal values and other meteorological data in the multi-source meteorological data of the target wind and light field station are cleaned, so that accurate cleaning and quality control of various observation data are realized, and the accuracy and consistency of the multi-source meteorological data are ensured.

[0024] In an optional implementation, the multi-source meteorological data of the target wind and light field station is subjected to abnormal value identification to obtain abnormal meteorological data, including:

[0025] The multi-source meteorological data of the target wind and light field station is subjected to abnormal identification by using a statistical analysis method, and the abnormal data is processed based on the abnormal identification result to obtain meteorological data after abnormal data processing.

[0026] The cleaned meteorological data is subjected to abnormal identification by using an artificial intelligence algorithm to obtain abnormal meteorological data.

[0027] The method provided by the embodiment is suitable for general topography, and the abnormal values in the multi-source meteorological data of the target wind and light field station are subjected to twice abnormal identification, so that accurate identification of the abnormal values is realized, and the accuracy of the meteorological data is ensured.

[0028] In an optional implementation, the multi-source meteorological data after data cleaning is subjected to sample expansion to obtain a virtual meteorological observation data set, including:

[0029] The multi-source meteorological data after data cleaning is subjected to feature extraction to obtain multi-source meteorological features.

[0030] The multi-source meteorological features are associated and fused to obtain the virtual meteorological observation data set.

[0031] The method provided by the embodiment is suitable for general topography, and the multi-source meteorological data after data cleaning and the associated fusion are realized, so that intelligent expansion of the limited sample of the wind and light field station micro area is realized, a high-resolution virtual observation data set of the regional kilometer level and the field station micro area hundred-meter level is formed, comprehensive monitoring of the meteorological condition is realized, and more accurate meteorological monitoring and prediction are provided for the field station.

[0032] In a second aspect, the present application provides a meteorological data optimization device suitable for general topography, which comprises:

[0033] The optimization module is configured to obtain an initial layout strategy of the wind and light field station, and to perform dynamic site selection optimization on the initial layout strategy of the wind and light field station to obtain an optimal layout strategy of the wind and light field station.

[0034] The data cleaning module is configured to obtain multi-source meteorological data of the target wind and light field station based on the optimal layout strategy of the wind and light field station, and clean the multi-source meteorological data of the target wind and light field station to obtain cleaned multi-source meteorological data.

[0035] The sample expansion module is configured to expand the cleaned multi-source meteorological data to obtain a virtual meteorological observation data set.

[0036] In a third aspect, the present application provides a computer device, comprising a memory and a processor, the memory and the processor are communicatively connected with each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the meteorological data optimization method for generalized terrain in the first aspect or any of the corresponding embodiments.

[0037] In a fourth aspect, the present application provides a computer readable storage medium, which stores computer instructions, and the computer instructions are used to make a computer execute the meteorological data optimization method for generalized terrain in the first aspect or any of the corresponding embodiments.

[0038] In a fifth aspect, the present application provides a computer program product, which comprises computer instructions, and the computer instructions are used to make a computer execute the meteorological data optimization method for generalized terrain in the first aspect or any of the corresponding embodiments. BRIEF DESCRIPTION OF DRAWINGS

[0039] In order to more clearly illustrate the specific embodiments of the present application or the technical solutions in the prior art, the drawings needed in the specific embodiments or prior art description will be briefly introduced as follows. Obviously, the drawings in the following description are some embodiments of the present application, and those skilled in the art can also obtain other drawings according to these drawings without creative labor.

[0040] Figure 1 is a flowchart of a meteorological data optimization method for generalized terrain according to an embodiment of the present application;

[0041] Figure 2 is a flowchart of another meteorological data optimization method for generalized terrain according to an embodiment of the present application;

[0042] Figure 3 is a flowchart of still another meteorological data optimization method for generalized terrain according to an embodiment of the present application;

[0043] Figure 4 is a flowchart of yet another meteorological data optimization method for generalized terrain according to an embodiment of the present application;

[0044] Figure 5 This is a structural block diagram of a meteorological data optimization device suitable for generalized terrain according to an embodiment of the present invention;

[0045] Figure 6 This is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] Due to the complex and diverse underlying surface and corresponding climate characteristics, a supply-demand imbalance exists between the efficiency of improving forecast accuracy and the rapidly growing field station operations. Firstly, field station meteorological data is relatively scarce, lacking high-quality, spatiotemporally continuous observation data. Secondly, due to the complexity of the natural environment and the limitations of monitoring equipment, the collected meteorological data often suffers from missing values, outliers, and noise. Furthermore, different data sources and formats cannot be standardized, resulting in inconsistent quality. A unified method for integrating multi-source meteorological data and efficiently utilizing field station data has not yet been established. Further efforts are needed to combine numerical models and artificial intelligence algorithms to continuously improve the spatiotemporal resolution of photovoltaic resources, forming field station meteorological data assets to provide data support for improving power forecast accuracy and enabling "meteorological+" data empowerment for micro-regional field stations. In addition, due to cost and environmental constraints, the actual number of sensors deployed is limited, leading to insufficient spatial resolution in the collected sample data. Therefore, quality control of station data and the efficient utilization of high-precision meteorological datasets formed by integrating multi-source data are crucial for improving the meteorological forecast accuracy of micro-regional field stations. This is essential for integrated demonstration and business applications across multiple scenarios and for achieving high-quality industrial development.

[0048] In summary, the relevant meteorological data collection methods have the following problems:

[0049] 1) Site selection for self-built monitoring stations in micro-areas of the station: When building self-built monitoring stations in micro-areas of the station, determining the optimal location and number of stations is a critical issue. Due to the lack of effective site selection theories and methods, the station layout may not be optimal, resulting in resource waste and insufficient monitoring data.

[0050] 2) Meteorological data quality control issues: There are problems such as missing values ​​and outliers in the raw meteorological observation data. These problems will affect the accuracy and reliability of the data, and thus affect the operation and maintenance decisions of the station and the forecast of power generation.

[0051] 3) Multi-source weather data fusion and sample expansion problem: Weather monitoring in the micro area of the station faces the problem of insufficient sample data spatial resolution. Due to cost and environmental restrictions, the number of sensors deployed is limited, which makes it impossible to obtain high-resolution weather data.

[0052] The embodiment of the application provides a weather data optimization method suitable for generalized terrain, optimizes station site selection through a conditional nonlinear optimal perturbation (CNOP) theoretical framework, realizes dynamic optimization and economic optimal configuration of self-built station layout, significantly improves weather data accuracy to improve station operation efficiency; secondly, develop weather data cleaning and quality control technology for wind and light stations, accurately clean and control data quality to ensure data accuracy and consistency, provide reliable data support for model training and prediction; finally, by fusing multi-source weather data and using intelligent expansion technology, a high-resolution virtual observation data set is formed, which enhances the feature expression and real-time monitoring capability of the data; not only improves the weather prediction accuracy of the station, but also promotes the sharing and application of data through standardized data receiving and transmitting interfaces, and provides strong technical support for high-quality development of the industry.

[0053] The embodiment of the application provides a weather data optimization method suitable for generalized terrain, and it should be noted that the execution subject of the weather data optimization method suitable for generalized terrain provided by the embodiment of the application can be a weather data optimization device suitable for generalized terrain. The weather data optimization device suitable for generalized terrain can be realized as part or all of an electronic device in the form of software, hardware or a combination of software and hardware, wherein the electronic device can be a server or a terminal, wherein the server in the embodiment of the application can be a server or a server cluster composed of multiple servers, and the terminal in the embodiment of the application can be a smart phone, a personal computer, a tablet computer, a wearable device, a smart robot and other smart hardware devices. In the following method embodiment, the execution subject is taken as an example to be explained.

[0054] According to the embodiment of the application, a weather data optimization method suitable for generalized terrain is provided, and it should be noted that the steps shown in the flowchart of the drawings can be executed in a computer system such as a group of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that here.

[0055] In the embodiment, a weather data optimization method suitable for generalized terrain is provided, which can be used in the above-mentioned electronic device, Figure 1is a flowchart of a meteorological data optimization method suitable for generalized terrain according to an embodiment of the present application, as shown in Figure 1 The flowchart includes the following steps:

[0056] In step S101, an initial layout strategy of a wind and light field station is obtained, and dynamic site selection optimization is performed on the initial layout strategy of the wind and light field station to obtain an optimal layout strategy of the wind and light field station.

[0057] Specifically, the efficient operation of the wind and light field station depends on accurate and timely meteorological data: in order to optimize the layout of the self-built station in the micro area of the field station, the conditional nonlinear optimal perturbation (CNOP) theory framework is adopted to perform dynamic site selection optimization on the station; wherein, the CNOP is a theory for analyzing and predicting the sensitivity of a complex system to initial condition perturbations, which can help identify the initial conditions that have the greatest impact on the system output, thereby guiding the site selection of the station.

[0058] In step S102, multi-source meteorological data of a target wind and light field station is obtained based on the optimal layout strategy of the wind and light field station, and data cleaning is performed on the multi-source meteorological data of the target wind and light field station to obtain cleaned multi-source meteorological data.

[0059] Specifically, for the meteorological data cleaning technology of the wind and light field station, it focuses on identifying and processing missing values and outliers, and through accurate cleaning and quality control, the accuracy and consistency of the observation data are ensured, providing high-quality data for subsequent model initial field, artificial intelligence training and correction.

[0060] In step S103, sample expansion is performed on the cleaned multi-source meteorological data to obtain a virtual meteorological observation data set.

[0061] Specifically, by fusing multi-source meteorological data such as field stations and regions, intelligent expansion technology is adopted to form a high-resolution virtual observation data set at the regional kilometer level and the field station micro area hundred-meter level.

[0062] The meteorological data optimization method suitable for generalized terrain provided in this embodiment realizes dynamic optimization of the layout of the self-built station by performing dynamic site selection optimization on the initial layout strategy of the wind and light field station, significantly improves the accuracy of the obtained meteorological data, and improves the operation efficiency of the wind and light field station; secondly, by performing data cleaning on the multi-source meteorological data of the target wind and light field station, the accuracy and consistency of the data are ensured, providing reliable data support for model training and prediction; finally, by performing sample expansion on the cleaned multi-source meteorological data, a high-resolution virtual meteorological observation data set is formed, the feature expression and real-time monitoring capability of the data are enhanced, and the meteorological prediction accuracy of the wind and light field station is improved.

[0063] The embodiment provides a meteorological data optimization method suitable for generalized terrain, which can be used for the electronic device, Figure 2 is a flowchart of a meteorological data optimization method suitable for generalized terrain according to an embodiment of the present application, as shown in the figure, the flow includes the following steps: Figure 2

[0064] Step S201, obtaining an initial layout strategy of a wind and light field station, performing dynamic site selection optimization on the initial layout strategy of the wind and light field station, and obtaining an optimal layout strategy of the wind and light field station.

[0065] Specifically, a conditional nonlinear optimal perturbation theory framework is used to perform dynamic site selection optimization on the self-built station in the micro region of the field station; by simulating and analyzing the sensitivity of different station numbers, positions and initial fields, the station layout scheme is optimized, the efficient use of the limited sample of the self-built station is realized, and the economic optimal configuration is achieved.

[0066] The above step S201 includes:

[0067] Step S2011, obtaining environmental parameters and economic configuration parameters of the wind and light field station based on the initial layout strategy of the wind and light field station, and constructing a comprehensive evaluation model based on the environmental parameters and economic configuration parameters of the wind and light field station.

[0068] Specifically, the environmental parameters such as the terrain and weather of the wind and light field station are collected, and the comprehensive evaluation model is constructed.

[0069] Further, based on remote sensing observation, the elevation terrain is obtained, and variables such as air temperature, precipitation and surface solar radiation output by the observation station are obtained, and then a comprehensive evaluation model is constructed based on correlation coefficient, root mean square error and other evaluation indexes; wherein the evaluation index coefficient = correlation coefficient + root mean square error, and the evaluation index coefficient is used to judge the overall performance index.

[0070] Further, the comprehensive evaluation model also considers economic factors, and minimizes the cost as much as possible under the premise of meeting the monitoring accuracy requirements.

[0071] Step S2012, evaluating multiple site selection schemes in the initial layout strategy of the wind and light field station by using the comprehensive evaluation model, and determining the optimal layout strategy of the wind and light field station based on the evaluation results.

[0072] Specifically, through simulation experiments, the influence of different station numbers, positions and initial field configurations on the model output is analyzed, the performance of each site selection scheme is evaluated, and it is further determined which positions of the station are most critical to improve the accuracy and efficiency of the meteorological monitoring of the entire field station micro region.

[0073] In some optional embodiments, the above step S2012 includes: ​

[0074] Step a1, determine the target function of the condition nonlinear optimal disturbance based on the comprehensive evaluation model.

[0075] Specifically, based on the comprehensive evaluation model, the key targets are determined, and the target function is established to represent the prediction uncertainty in the prediction period, that is, the core target is defined: the core index to be optimized for station site selection, for example: "error minimization", "sensitive area coverage maximization", or "cost-precision balance optimization" of the field station micro area meteorological element (temperature, wind speed, humidity, etc.) monitoring.

[0076] Further, the system boundary is drawn: the geographical range of the field station micro area (such as the terrain undulating area, building dense area, etc.) is determined, and the key influencing factors of the meteorological system (such as terrain height, surface roughness, underlying surface type, etc.) are determined, which define the range for subsequent model construction, that is, the constraint conditions of the above target function include: the geographical range of the candidate site, the upper limit of the number of new sites, and the physical limit of the initial disturbance amplitude (representing the observation error).

[0077] Further, the basic data of the field station is collected, including: terrain data (elevation, slope, slope direction); historical meteorological data (long-term observed temperature, wind speed, precipitation, etc. time series data); underlying surface data (vegetation coverage, building distribution, etc.); based on the above data, a nonlinear numerical model describing the evolution of meteorological elements in the field station micro area is constructed, which can reflect the nonlinear relationship between "station observation data (initial conditions)" and "meteorological field simulation results (system output)".

[0078] Further, the station layout is regarded as a "disturbance" to the initial conditions of the system, which specifically includes: changes in station location (such as shifts in coordinates (x, y)), changes in the number of stations (such as increasing / decreasing n stations), and changes in station observation accuracy (such as fluctuations in sensor error), and the above constraints are set.

[0079] Further, based on the CNOP theory, the goal is to find the disturbance that has the greatest impact on the system output (i.e. the station layout that can most reduce the monitoring error), and the function form can be represented as: J (disturbance) = maximum of system output error (or "maximum monitoring accuracy of sensitive area"); wherein "system output error" refers to the deviation between the simulation results of the meteorological model and the true meteorological field (or high-precision reference data) when the station layout is disturbed; the larger the target function value, the more significant the impact of the disturbance on the system output, that is, the layout can more sensitively capture the key changes of the meteorological field, and it is the core station location that needs to be preserved or optimized.

[0080] Step a2, numerically solve the target function to obtain the optimal layout strategy of the wind and light station.

[0081] Specifically, the space maximum value area of the CNOP solution is taken as a new station site location, observation equipment is deployed and data is assimilated; as the prediction demand evolves or the network expands, the above steps are repeated: based on the updated background field and the prediction target, the CNOP is recalculated to identify the sensitive area in the next stage, the station layout is continuously optimized, a dynamic closed loop of “evaluation-siting-assimilation-re-evaluation” is formed, and the robustness of the target prediction is gradually improved.

[0082] Further, the disturbance of “maximizing the objective function J under the constraint condition” (i.e., the optimal station layout) is solved by a numerical method, specifically including:

[0083] 1) Initialization: A set of initial station layout schemes (such as randomly generating a number of candidate positions) is set.

[0084] 2) Iterative optimization: For each candidate layout scheme, the corresponding meteorological field output is simulated by a meteorological numerical model; the objective function value J (i.e., the degree of influence on the system output) of the scheme is calculated; an optimization algorithm (such as the adjoint method, genetic algorithm, simulated annealing method) is used to iteratively update the station layout, gradually approaching the optimal solution (CNOP solution) of “J maximum”, wherein, due to the nonlinearity of the meteorological system, multiple nonlinear iterations are required to avoid falling into a local optimum, ensuring that the globally optimal station layout is found.

[0085] 3) Key output: A set of “CNOP key station positions” is obtained, i.e., the station at this position has the greatest impact on the monitoring accuracy of the meteorological field, and is the core node of layout optimization.

[0086] 4) Multi-scheme evaluation and dynamic adjustment: Performance evaluation: based on the CNOP solution, a plurality of candidate station layout schemes (such as combinations containing different numbers of key stations) are generated, and the accuracy indicators (meteorological field simulation error and sensitive area coverage) and economic indicators (construction cost and maintenance cost) of each scheme are evaluated.

[0087] 5) Dynamic optimization: combined with the changes in the station operation demand (such as the addition of sensitive areas or budget adjustments), the latest topography and meteorological data are regularly collected (such as every quarter / year), the meteorological numerical model is updated, the CNOP is recalculated, and the station layout is adjusted (such as adding / removing stations, changing positions), to achieve “dynamic adaptation”.

[0088] Step S202, based on the optimal layout strategy of the wind and light station station, the multi-source meteorological data of the target wind and light station is obtained, and the multi-source meteorological data after data cleaning is obtained. For details, please refer to Figure 1 The step S102 of the embodiment shown in the figure will not be repeated here.

[0089] Step S203, sample expansion is performed on the multi-source meteorological data after data cleaning to obtain a virtual meteorological observation data set. For details, please refer to Figure 1 Step S103 of the embodiment shown will not be described here again.

[0090] The meteorological data optimization method suitable for generalized terrain provided in this embodiment can efficiently utilize limited observation samples through dynamic site selection optimization, realize economic optimal configuration of station layout, improve the operation efficiency of the station, and reduce the construction and maintenance cost. The target function of the conditional nonlinear optimal disturbance is determined, and the target function of the conditional nonlinear optimal disturbance is solved. The site selection optimization problem of the self-built station in the micro area of the station is studied by using the conditional nonlinear optimal disturbance theory, and efficient utilization and economic optimization of the limited samples of the self-built station are realized.

[0091] In this embodiment, a meteorological data optimization method suitable for generalized terrain is provided, which can be used in the electronic device described above, Figure 3 is a flowchart of a meteorological data optimization method suitable for generalized terrain according to an embodiment of the present application, as shown in Figure 3 The flowchart includes the following steps:

[0092] Step S301, an initial layout strategy of a wind and light station is obtained, and dynamic site selection optimization is performed on the initial layout strategy of the wind and light station to obtain an optimal layout strategy of the wind and light station. For details, please refer to Figure 2 Step S201 of the embodiment shown will not be described here again.

[0093] Step S302, multi-source meteorological data of a target wind and light station is obtained based on the optimal layout strategy of the wind and light station, and data cleaning is performed on the multi-source meteorological data of the target wind and light station to obtain multi-source meteorological data after data cleaning.

[0094] Specifically, the above step S302 includes:

[0095] Step S3021, the target wind and light station is selected based on the optimal layout strategy of the wind and light station, and the multi-source meteorological data of the target wind and light station is obtained.

[0096] Specifically, the optimal site of the wind and light station is determined according to the optimal layout strategy of the wind and light station, and the multi-source meteorological data of the wind and light station at the optimal site is collected, wherein the multi-source meteorological data includes observation data of a ground meteorological station, large-scale meteorological information provided by satellite remote sensing, accurate precipitation data observed by radar, and vertical atmospheric data obtained by a sounding balloon.

[0097] Step S3022, the multi-source meteorological data of the target wind and light station is preprocessed to obtain preprocessed meteorological data.

[0098] Specifically, the multi-source meteorological data of the target wind and light field station is preprocessed, including data format unification, time series synchronization, and removal of repeated records.

[0099] Further, a standardized meteorological data receiving and transmitting interface is developed, which follows an open and universal data exchange standard, so that data of different sources and formats can be uniformly processed and analyzed. The establishment of the standardized interface not only improves the efficiency of data processing, but also reduces the complexity of system integration, so that meteorological data can be more easily applied to the operation and management of the field station.

[0100] Step S3023, abnormal value identification is performed on the multi-source meteorological data of the target wind and light field station to obtain abnormal meteorological data.

[0101] In some optional embodiments, the above step S3023 includes:

[0102] Step b1, using statistical analysis method to identify abnormal data of the target wind and light field station, based on the abnormal identification result to process abnormal data, to obtain abnormal data processed meteorological data.

[0103] Specifically, statistical analysis methods such as mean, standard deviation, and quartile distance are used to identify abnormal values in the data. For the identified abnormal values, different processing strategies are taken according to their characteristics, such as using the average value of adjacent time points, using historical data interpolation, or correcting according to the physical model.

[0104] Further, different processing strategies include:

[0105] 1) Threshold check (range control): According to the physical possibility and climate extreme value, set the reasonable upper and lower limit of each element (temperature, pressure, humidity, wind, etc.), data exceeding the threshold is considered as error or suspicious, directly marked or removed.

[0106] 2) Internal consistency check (logic verification): Check whether the same site, different elements at the same time meet the physical logic relationship; for example: dew point temperature should not be higher than air temperature; when precipitation occurs, relative humidity should be close to saturation.

[0107] 3) Spatio-temporal consistency check (background field verification): Use the continuity of data in time or space for verification, check the reasonable change of adjacent time data in time series (such as temperature short-time dramatic change needs weather basis); compare the difference and gradient of the observation value of the target station and its surrounding stations in space, significantly deviate from the regional background field, then mark as abnormal.

[0108] Step b2, using artificial intelligence algorithm to identify abnormal data of the cleaned meteorological data, to obtain abnormal meteorological data.

[0109] Specifically, artificial intelligence algorithms such as random forest, XGBoost (eXtreme Gradient Boosting, optimized distributed gradient boosting library), and generative adversarial network are introduced to automatically identify and process complex abnormal patterns. Artificial intelligence algorithms can learn from sample data (i.e., meteorological data after abnormal data processing) to improve the accuracy of missing or abnormal value expansion, ensure data accuracy and consistency through accurate data cleaning and quality control, and provide effective support for subsequent data analysis, physical model initialization, and artificial intelligence training.

[0110] Step S3024, the abnormal meteorological data is corrected to obtain the multi-source meteorological data after data cleaning.

[0111] Specifically, the missing values in the abnormal meteorological data are interpolated, and the abnormal values in the abnormal meteorological data are deleted or replaced with the average value of the adjacent time points.

[0112] Step S303, sample expansion is performed on the multi-source meteorological data after data cleaning to obtain a virtual meteorological observation data set. For details, please refer to Figure 2 Step S203 of the embodiment shown in

[0113] The meteorological data optimization method provided in this embodiment is suitable for generalizing terrain, and through cleaning of missing values, abnormal values and other meteorological data in the multi-source meteorological data of the target wind and light field station, accurate cleaning and quality control of various observation data are realized to ensure the accuracy and consistency of the multi-source meteorological data.

[0114] In this embodiment, a meteorological data optimization method suitable for generalizing terrain is provided, which can be used in the electronic device described above, Figure 4 is a flowchart of a meteorological data optimization method suitable for generalizing terrain according to an embodiment of the present application, as Figure 4 shown, the flowchart includes the following steps:

[0115] Step S401, an initial layout strategy of a wind and light field station is obtained, and the initial layout strategy of the wind and light field station is dynamically addressed and optimized to obtain an optimal layout strategy of the wind and light field station. For details, please refer to Figure 3 Step S301 of the embodiment shown in

[0116] Step S402, based on the optimal layout strategy of the wind and light field station, the multi-source meteorological data of the target wind and light field station is obtained, and the multi-source meteorological data of the target wind and light field station is cleaned to obtain the multi-source meteorological data after data cleaning. For details, please refer to Figure 3 Step S302 of the embodiment shown in

[0117] In step S403, the multi-source weather data after data cleaning is sample expanded to obtain a virtual weather observation data set.

[0118] Specifically, the step S403 includes:

[0119] In step S4031, the multi-source weather data after data cleaning is feature extracted to obtain multi-source weather features.

[0120] Specifically, the heterogeneous raw data of different sources (such as satellite, radar, and ground station) and types (image, time series, and point data) is cleaned, standardized, and spatio-temporally aligned; key features (such as cloud texture and temperature trend) are automatically extracted by using convolutional neural networks, autoencoders, and the like; and the multi-source weather data after data cleaning is converted into unified and fusable deep feature representation, thereby laying a foundation for subsequent fusion.

[0121] In step S4032, the multi-source weather features are associated and fused to obtain a virtual weather observation data set.

[0122] Specifically, a specific fusion architecture (such as an attention mechanism and a graph neural network) is used to model the complex nonlinear relationship between multi-source features, that is, for the complex relationship (such as spatial dependence and cross-modal association) of multi-source features, an advanced architecture such as an attention mechanism and a graph neural network is used to dynamically weigh the importance of different data sources and model the internal relationship, generate joint features that can comprehensively reflect the target state, and break through the limitations of a single data source; for example: the importance of different data sources is dynamically allocated through attention weights; the spatial dependence between stations is modeled by using a graph neural network (GNN); and cross-modal association is learned by using a multi-modal transformer to generate joint features that can comprehensively reflect the target state.

[0123] Specifically, the step of associating and fusing the multi-source weather features includes:

[0124] 1) Association type identification: spatial association: there is a spatial spillover effect between the temperature features of adjacent weather stations; temporal association: there is a time series dependence between the radar echo features of the previous 1 hour and the precipitation features of the current hour; cross-modal association: there is a nonlinear mapping relationship between the texture features (image modal) of the satellite cloud image and the humidity features (numerical modal) of the ground station.

[0125] 2) Association strength quantification:

[0126] Statistical method: the Pearson correlation coefficient, mutual information, and the like are used to quantify the linear or nonlinear association strength;

[0127] Initial model: use simple network (such as multi-layer perception MLP) to fit the mapping relationship between multi-source features, and judge the correlation complexity through loss value (the lower the loss, the more significant the correlation).

[0128] 3) Fusion architecture: according to the correlation mode analysis result, select or design the appropriate deep fusion architecture to model the complex relationship between features; the fusion architecture and the applicable scenario are as follows:

[0129] A, attention mechanism: applicable scenario: dynamic change of feature correlation strength (such as different importance of satellite data and ground data at different times); operation: automatically allocate the contribution of each source feature by learning "attention weight" (such as giving high weight to strong correlation features), for example, in weather forecasting, use spatial attention to focus on the features of the surrounding stations that have the greatest impact on the target area, and use time attention to highlight the radar data in the key period.

[0130] B, graph neural network: applicable scenario: features with explicit topological relationship (such as geographical distance between stations, connection relationship of sensor network); operation: abstract features as "nodes", use edge weight to represent correlation strength, pass node information through graph convolution, graph attention, etc. Layer to model spatial or structural dependence, for example, use GNN to fuse wind speed features of different wind farm stations to capture the airflow propagation correlation between stations.

[0131] C, multi-modal Transformer: applicable scenario: complex interaction of cross-modal features (such as text, image, numerical value) (such as the correlation between satellite images and weather report text); operation: learn feature dependence within the modality through self-attention mechanism, model inter-modal correlation through cross-attention mechanism, generate fused context features, for example, in disaster identification, use Transformer to fuse satellite image features (visual modality) and disaster description text features (language modality), to improve recognition accuracy.

[0132] D, hybrid architecture: applicable scenario: multiple types of correlation coexist (such as containing spatial, temporal, and cross-modal correlation at the same time); operation: combine the above architectures, such as "GNN (spatial correlation) + LSTM (temporal correlation) + attention (dynamic weight)", model feature dependence in all directions.

[0133] 4) Joint feature generation: operate multi-source features through designed fusion architecture, output "joint features" containing comprehensive information, generation methods include: weighted fusion: weighted sum of multi-source features based on attention weight or preset rules (such as correlation strength); splicing fusion: directly splice multi-source features in dimension (such as splicing image features and time sequence features into higher dimensional vectors), and then compress into compact joint features through fully connected layer or convolution layer; interactive fusion: strengthen the interaction between features through element-level product, matrix multiplication and other operations; hierarchical fusion: first locally fuse features of the same type (such as fusing multiple satellite image features into "image joint features"), and then globally fuse different types of local joint features to reduce computational complexity.

[0134] 5) Model training and optimization: train the fusion model and optimize the parameters in the direction of task objectives (such as minimizing prediction error and maximizing classification accuracy).

[0135] Further, the fused joint features are input into a task-driven model to generate final output (such as weather forecast and disaster identification); wherein the task-driven model includes LSTM (Long Short-Term Memory) prediction, ResNet (Residual Network) classification, etc.

[0136] The meteorological data optimization method provided in the embodiment is suitable for generalized terrain, and can realize effective expansion of limited samples of wind and light field stations in micro regions, form high-resolution virtual observation data sets of regional kilometers and micro-regional hundreds of meters of stations, and realize comprehensive monitoring of meteorological conditions, thereby providing more accurate meteorological monitoring and prediction for stations.

[0137] In the embodiment, a meteorological data optimization device suitable for generalized terrain is also provided, which is used to implement the above-mentioned embodiments and preferred embodiments, and details are not repeated. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware, or a combination of software and hardware is also possible and is contemplated.

[0138] The embodiment provides a meteorological data optimization device suitable for generalized terrain, as shown in Figure 5 , comprising:

[0139] The optimization module 501 is configured to obtain an initial layout strategy of the wind and light field station, and perform dynamic site selection optimization on the initial layout strategy of the wind and light field station to obtain an optimal layout strategy of the wind and light field station.

[0140] The data cleaning module 502 is configured to obtain multi-source meteorological data of a target wind and light field station based on the optimal layout strategy of the wind and light field station, and perform data cleaning on the multi-source meteorological data of the target wind and light field station to obtain cleaned multi-source meteorological data.

[0141] The sample expansion module 503 is configured to perform sample expansion on the cleaned multi-source meteorological data to obtain a virtual meteorological observation data set.

[0142] In some optional embodiments, the optimization module 501 comprises:

[0143] The construction unit is configured to obtain environmental parameters and economic configuration parameters of the wind and light field station based on the initial layout strategy of the wind and light field station, and construct a comprehensive evaluation model based on the environmental parameters and the economic configuration parameters of the wind and light field station.

[0144] The evaluation unit is configured to evaluate a plurality of site selection schemes in the initial layout strategy of the wind and light field station by using the comprehensive evaluation model, and determine the optimal layout strategy of the wind and light field station based on an evaluation result.

[0145] In some optional embodiments, the evaluation unit comprises:

[0146] The determination sub-unit is configured to determine a target function of a conditional nonlinear optimal disturbance based on the comprehensive evaluation model.

[0147] The solving sub-unit is configured to perform numerical solution on the target function to obtain the optimal layout strategy of the wind and light field station.

[0148] In some optional embodiments, the data cleaning module 502 comprises:

[0149] The obtaining unit is configured to select a target wind and light field station based on the optimal layout strategy of the wind and light field station, and obtain multi-source meteorological data of the target wind and light field station.

[0150] The preprocessing unit is configured to perform preprocessing on the multi-source meteorological data of the target wind and light field station to obtain preprocessed meteorological data.

[0151] The identification unit is configured to perform outlier identification on the multi-source meteorological data of the target wind and light field station to obtain abnormal meteorological data.

[0152] The correction unit is configured to correct the abnormal meteorological data to obtain the cleaned multi-source meteorological data.

[0153] In some optional embodiments, the identifying unit comprises:

[0154] The first identifying sub-unit is configured to perform abnormality identification on the multi-source meteorological data of the target wind and light field station by using a statistical analysis method, perform abnormal data processing based on the abnormality identification result, and obtain meteorological data after abnormal data processing.

[0155] The second identifying sub-unit is configured to perform abnormality identification on the cleaned meteorological data by using an artificial intelligence algorithm, and obtain abnormal meteorological data.

[0156] In some optional embodiments, the sample expansion module 503 comprises:

[0157] The extracting unit is configured to perform feature extraction on the multi-source meteorological data after data cleaning, and obtain multi-source meteorological features.

[0158] The correlation fusion unit is configured to perform correlation fusion on the multi-source meteorological features, and obtain a virtual meteorological observation data set.

[0159] Further function descriptions of the above-mentioned various modules and units are the same as those of the above-mentioned corresponding embodiments, and will not be repeated here.

[0160] The meteorological data optimization device suitable for generalized terrain in the embodiment is presented in the form of a functional unit. The unit herein refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and a memory executing one or more software or fixed programs, and / or other devices that can provide the above-mentioned functions.

[0161] The embodiment of the present application also provides a computer device having the above-mentioned Figure 5 meteorological data optimization device suitable for generalized terrain.

[0162] Please refer to Figure 6 , Figure 6 is a structural schematic diagram of a computer device provided by an optional embodiment of the present application, as Figure 6As shown, the computer device includes one or more processors 10, memory 20, and interfaces 50 for external devices such as modems and network interfaces. The one or more processors 10 can be implemented as one or more central processing units (CPUs), one or more microprocessors, microcontrollers, digital signal processors, specialized processors or controller, or one or more processors of any equivalent known in the art. In some embodiments, the one or more processors 10 can be implemented as a combination of one or more of the above physical processors and / or one or more software or firmware modules. The software or firmware can be stored in a non-transitory computer readable medium such as memory 20, a storage device or any equivalent medium known in the art. Figure 6 The processor 10 is taken as an example in the embodiments.

[0163] The processor 10 can be a central processing unit, a network processing unit or a combination thereof. The processor 10 can further include a hardware chip. The hardware chip can be an application specific integrated circuit, a programmable logic device or a combination thereof. The programmable logic device can be a complex programmable logic device, a field programmable logic device, a generic array logic or any combination thereof.

[0164] The memory 20 stores instructions that are executable by the at least one processor 10, so as to enable the at least one processor 10 to perform the method shown in the above embodiments.

[0165] The memory 20 can include a program storage area and a data storage area. The program storage area can store an operating system, application programs required by at least one function, and the like. The data storage area can store data created according to the use of the computer device, and the like. In addition, the memory 20 can include a high-speed random access memory, and can further include a non-transitory memory such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state memory device. In some alternative embodiments, the memory 20 can optionally include a memory disposed remotely from the processor 10, and these remote memories can be connected to the computer device through a network. Examples of the network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0166] The memory 20 can include a volatile memory such as a random access memory, and can also include a non-volatile memory such as a flash memory, a hard disk or a solid state disk, and can further include a combination of the above kinds of memories.

[0167] The computer device further includes an input device 30 and an output device 40. The processor 10, the memory 20, the input device 30 and the output device 40 can be connected by a bus or other means,Figure 6 The bus connection is taken as an example.

[0168] The input device 30 can receive inputted digital or character information, and generate key signal inputs related to user settings and function controls of the computer device, such as a touch screen, a keypad, a mouse, a trackpad, a touchpad, a pointing stick, one or more mouse buttons, a trackball, a joystick, etc. The output device 40 can include a display device, an auxiliary lighting device (e.g., an LED), a tactile feedback device (e.g., a vibration motor), etc. The display device includes, but is not limited to, a liquid crystal display, a light-emitting diode, a display, and a plasma display. In some alternative embodiments, the display device can be a touch screen.

[0169] The embodiments of the present application also provide a computer readable storage medium, the method according to the embodiments of the present application can be implemented in hardware, firmware, or recorded in a storage medium, or stored in a remote storage medium or a non-transitory machine readable storage medium and downloaded from a network and stored in a local storage medium, so that the method described herein can be processed by such software on a storage medium using a general purpose computer, a special purpose processor, or programmable or special purpose hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only memory, a random access memory, a flash memory, a hard disk or a solid state disk, etc.; further, the storage medium can also include a combination of the above types of memories. It can be understood that the computer, processor, microprocessor controller or programmable hardware includes a storage component that can store or receive software or computer code, when the software or computer code is accessed and executed by the computer, processor or hardware, the method shown in the above embodiments is implemented.

[0170] Part of the present application can be applied as a computer program product, for example, computer program instructions, when executed by a computer, through the operation of the computer, the method and / or technical solutions according to the present application can be called or provided. Those skilled in the art should understand that the form of computer program instructions in computer readable medium includes but is not limited to source file, executable file, installation package file, etc., and accordingly, the way of computer program instructions executed by computer includes but is not limited to: the computer directly executes the instructions, or the computer compiles the instructions and then executes the corresponding compiled program, or the computer reads and executes the instructions, or the computer reads and installs the instructions and then executes the corresponding installed program. Here, the computer readable medium can be any available computer readable storage medium or communication medium accessible to the computer.

[0171] While embodiments of the application have been described in connection with the preferred embodiments of the various figures, those of ordinary skill in the art will appreciate that various modifications and variations of the preferred embodiments can be employed without departing from the spirit and scope of the application.

Claims

1. A method for optimizing meteorological data suitable for generalized terrain, characterized in that, The method comprises: acquiring an initial layout strategy of a wind and light field station, and performing dynamic site selection optimization on the initial layout strategy of the wind and light field station to obtain an optimal layout strategy of the wind and light field station; based on the optimal layout strategy of the wind and light field station, acquiring multi-source meteorological data of a target wind and light field station, and performing data cleaning on the multi-source meteorological data of the target wind and light field station to obtain cleaned multi-source meteorological data; performing sample expansion on the cleaned multi-source meteorological data to obtain a virtual meteorological observation data set.

2. The method of claim 1, wherein, The dynamic site selection optimization on the initial layout strategy of the wind and light field station to obtain the optimal layout strategy of the wind and light field station comprises: based on the initial layout strategy of the wind and light field station, obtaining environmental parameters and economic configuration parameters of the wind and light field station, and constructing a comprehensive evaluation model based on the environmental parameters and the economic configuration parameters of the wind and light field station; using the comprehensive evaluation model to evaluate a plurality of site selection schemes in the initial layout strategy of the wind and light field station, and determining the optimal layout strategy of the wind and light field station based on the evaluation results.

3. The method of claim 2, wherein, The evaluation of the plurality of site selection schemes in the initial layout strategy of the wind and light field station using the comprehensive evaluation model and the determination of the optimal layout strategy of the wind and light field station based on the evaluation results comprise: determining a target function of conditional nonlinear optimal disturbance based on the comprehensive evaluation model; numerically solving the target function to obtain the optimal layout strategy of the wind and light field station.

4. The method of claim 1, wherein, The acquisition of the multi-source meteorological data of the target wind and light field station based on the optimal layout strategy of the wind and light field station, the data cleaning on the multi-source meteorological data of the target wind and light field station, and the obtaining of the cleaned multi-source meteorological data comprise: selecting a target wind and light field station based on the optimal layout strategy of the wind and light field station, and acquiring multi-source meteorological data of the target wind and light field station; preprocessing the multi-source meteorological data of the target wind and light field station to obtain preprocessed meteorological data; identifying abnormal values in the multi-source meteorological data of the target wind and light field station to obtain abnormal meteorological data; correcting the abnormal meteorological data to obtain the cleaned multi-source meteorological data.

5. The method of claim 4, wherein, The identification of abnormal values in the multi-source meteorological data of the target wind and light field station to obtain abnormal meteorological data comprises: using a statistical analysis method to identify abnormalities in the multi-source meteorological data of the target wind and light field station, and performing abnormal data processing based on the abnormal identification results to obtain meteorological data after abnormal data processing; using an artificial intelligence algorithm to identify abnormalities in the cleaned meteorological data to obtain the abnormal meteorological data.

6. The method of claim 1, wherein, The sample expansion on the cleaned multi-source meteorological data to obtain a virtual meteorological observation data set comprises: extracting features from the cleaned multi-source meteorological data to obtain multi-source meteorological features; correlating and fusing the multi-source meteorological features to obtain the virtual meteorological observation data set.

7. A weather data optimization device suitable for generalized terrain, characterized by, The device comprises: An optimization module is configured to obtain an initial layout strategy of wind-solar farm stations, and perform dynamic site selection optimization on the initial layout strategy of the wind-solar farm stations to obtain an optimal layout strategy of the wind-solar farm stations. A data cleaning module is configured to obtain multi-source meteorological data of a target wind-solar farm based on the optimal layout strategy of the wind-solar farm stations, and perform data cleaning on the multi-source meteorological data of the target wind-solar farm to obtain cleaned multi-source meteorological data. A sample expansion module is configured to perform sample expansion on the cleaned multi-source meteorological data to obtain a virtual meteorological observation data set.

8. A computer device, comprising: The application discloses a meteorological data optimization method suitable for generalized terrain, and relates to the technical field of meteorological data optimization. The computer readable storage medium stores computer instructions, and the computer instructions are used to make the computer execute the meteorological data optimization method suitable for generalized terrain.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer instructions, and the computer instructions are used to make the computer execute the meteorological data optimization method suitable for generalized terrain.

10. A computer program product, characterised in that, The computer readable storage medium stores computer instructions, and the computer instructions are used to make the computer execute the meteorological data optimization method suitable for generalized terrain.

Images