A wind farm key meteorological factor prediction method and system based on multi-source data

Abstract

The application discloses a kind of based on multi-source data's wind farm key meteorological factor prediction method and system, it is related to intelligent data processing technical field, the method includes: connecting meteorological platform obtains historical wind farm meteorological factor data, carries out multidimensional data source analysis, determines A data source, B meteorological characteristic element;B meteorological characteristic element, A data source are input into meteorological characteristic prediction model and obtain wind farm meteorological characteristic information, identification is carried out in combination with historical wind farm meteorological data to multilevel meteorological factor response interval, generate multilevel meteorological factor sensitive index, then in combination with the meteorological factor prediction data collected, the key meteorological factor of target wind farm is forecasted.The application solves the technical problem that the accuracy of the prediction result caused by the single data source of wind farm weather forecast in the prior art is low, realizes the technical effect of improving the accuracy of wind farm key meteorological factor prediction result.

Description

A method and system for forecasting key meteorological factors of wind farms based on multi-source data Technical Field

[0001] This invention relates to the field of intelligent data processing technology for wind farms, specifically to a method and system for forecasting key meteorological factors of wind farms based on multi-source data. Background Technology

[0002] To mitigate global warming, the call for clean energy is growing louder. Wind energy, as an important clean energy source, has been widely developed in many parts of the world, with an increasing number of large-scale wind farms being built and put into operation. The construction of wind farms pays close attention to the impact of abnormal weather and meteorological events, making weather forecasting for wind farms increasingly important.

[0003] Current wind farm weather forecast data is mostly derived from single atmospheric models or multi-model integrated forecasts without correction for measured data. This results in significant discrepancies with actual observations, only meeting the requirements for precision, and requiring further improvement in accuracy. Therefore, existing technologies suffer from the technical problem of low accuracy in wind farm weather forecasts due to a single data source. Summary of the Invention

[0004] This application provides a method and system for forecasting key meteorological factors of wind farms based on multi-source data, which solves the technical problem of low accuracy of meteorological forecast results caused by a single meteorological data source in the prior art.

[0005] The first aspect of this application provides a method for forecasting key meteorological factors of wind farms based on multi-source data, the method comprising:

[0006] By connecting to the meteorological platform, historical meteorological data of wind farms can be obtained;

[0007] The historical meteorological data of the wind farms are analyzed to obtain high-impact weather information for the historical wind farms;

[0008] Multidimensional data source analysis was performed on the historical high-impact weather information of the wind farm to determine A data sources, where A is a positive integer greater than 2;

[0009] Based on the A data sources, feature pairing is performed in the meteorological feature pairing library to determine B meteorological feature elements. The B meteorological feature elements and the A data sources are then input into the meteorological law layer of the meteorological feature prediction model to output the meteorological feature information of the wind farm. Here, B is a positive integer greater than 1.

[0010] Based on the historical meteorological data of the wind farm and the meteorological characteristic information of the wind farm, the response intervals of multi-level meteorological factors are identified, and a multi-level meteorological factor sensitivity index is generated based on the identification results.

[0011] Meteorological factor prediction data are collected, and the weather forecast for the target wind farm is made based on the meteorological factor prediction data and the multi-level meteorological factor sensitivity index.

[0012] A second aspect of this application provides a remote intelligent control system for a hydraulic gate, the system comprising:

[0013] A meteorological data acquisition module is used to acquire historical meteorological data of wind farms by connecting to a meteorological platform.

[0014] The data analysis module is used to analyze the historical meteorological data of the wind farm to obtain high-impact weather information of the historical wind farm.

[0015] A multidimensional data source analysis module is used to perform multidimensional data source analysis on the historical high-impact weather information of wind farms to determine A data sources.

[0016] The meteorological characteristic information acquisition module is used to perform feature matching of the meteorological characteristic matching library based on the A data sources, determine B meteorological characteristic elements, input the B meteorological characteristic elements and the A data sources into the meteorological law layer in the meteorological characteristic prediction model, and output the meteorological characteristic information of the wind farm, where B is a positive integer greater than 1.

[0017] The multi-level meteorological factor sensitivity index generation module, wherein the multi-level meteorological factor response interval identification module is used to identify the multi-level meteorological factor response interval based on the historical wind farm meteorological data and the wind farm meteorological characteristic information, and generate a multi-level meteorological factor sensitivity index based on the identification results;

[0018] The weather forecast module is used to collect meteorological factor prediction data and forecast the weather of the target wind farm based on the meteorological factor prediction data and the multi-level meteorological factor sensitivity index.

[0019] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0020] This application provides a method for forecasting key meteorological factors of wind farms based on multi-source data, which relates to the field of intelligent data processing technology. It solves the technical problem of low accuracy of meteorological forecast results caused by a single data source in the existing technology for wind farm meteorological forecasts, and realizes the technical effect of improving the accuracy of meteorological forecast results by performing multi-source data meteorological forecasts for wind farms. Attached Figure Description

[0021] 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.

[0022] Figure 1 is a schematic flowchart of a method for forecasting key meteorological factors of wind farms based on multi-source data provided in an embodiment of this application.

[0023] Figure 2 is a flowchart illustrating the process of determining A data sources in a wind farm key meteorological factor forecasting method based on multi-source data provided in an embodiment of this application.

[0024] Figure 3 is a flowchart illustrating the output of wind farm meteorological characteristic information in a wind farm key meteorological factor forecasting method based on multi-source data provided in an embodiment of this application.

[0025] Figure 4 is a schematic diagram of the structure of a wind farm key meteorological factor forecasting system based on multi-source data provided in an embodiment of this application.

[0026] Figure labeling: Meteorological data acquisition module 11, data analysis module 12, multidimensional data source analysis module 13, meteorological characteristic information acquisition module 14, multi-level meteorological factor sensitivity index generation module 15, weather forecast module 16. Detailed Implementation

[0027] This application provides a method for forecasting key meteorological factors of wind farms based on multi-source data, which is used to address the technical problem of low accuracy of meteorological forecast results caused by a single data source for wind farm meteorological forecasts in the prior art.

[0028] 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.

[0029] It should be noted that the terms "first," "second," etc., in the specification, claims, 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 comprises 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.

[0030] Example 1

[0031] As shown in Figure 1, this application provides a method for forecasting key meteorological factors of wind farms based on multi-source data. The method includes:

[0032] S100: Obtain historical meteorological data from wind farms by connecting to a meteorological platform;

[0033] Specifically, the meteorological platform refers to a meteorological data sharing platform, which utilizes information technology to collect, analyze, store, share, and use meteorological data, effectively supporting various meteorological operations. Wind power refers to wind-generated electricity, and a wind farm is a wind power generation base. Since the meteorological types encompassed by a target wind farm are within a certain range, it is necessary to collect historical meteorological data from the target wind farm. This data is then aggregated to represent the meteorological data existing at the target wind farm up to the current moment, resulting in the historical wind farm meteorological data. This historical wind farm meteorological data is time-stamped and corresponds one-to-one with the historical wind farm meteorological data. By logging into the meteorological data sharing platform and searching for this historical wind farm meteorological data, it can be obtained and used as the basis for subsequent meteorological analysis.

[0034] S200: Analyze the historical meteorological data of the wind farm to obtain high-impact weather information of the historical wind farm;

[0035] Specifically, analyzing the historical wind farm meteorological data involves screening, classifying, and organizing the meteorological data existing at the wind farm up to the current moment. High-impact weather for wind farms refers to weather events such as typhoons, low temperatures, icing, thunderstorms, and sandstorms that affect the safe operation of wind farms. Since wind farm construction focuses heavily on the impact of abnormal weather and meteorological events, high-impact weather information for wind farms is selected as the basic data for meteorological forecasting. From the historical wind farm meteorological data, weather conditions such as typhoons, low temperatures, icing, thunderstorms, and sandstorms that affect the safe operation of wind farms are screened out, and weather data under those conditions is obtained, including images, historical time periods, atmospheric conditions, air pressure, humidity, temperature, and weather distribution. This data constitutes the historical high-impact weather information for wind farms and can serve as the basic data source for subsequent meteorological analysis.

[0036] S300: Perform multi-dimensional data source analysis on the historical high-impact weather information of the wind farm to determine A data sources, where A is a positive integer greater than 2;

[0037] Specifically, multidimensional data source analysis refers to selecting data from different dimensions based on the historical high-impact weather information of wind farms, and conducting analysis on these different dimensions to obtain data sources of different dimensions. For example, based on the historical high-impact weather information of wind farms, real-time image information of the target wind farm can be selected. By analyzing the brightness of the real-time image, the light intensity, visibility, cloud thickness, etc., of the environment at that time can be determined. Based on the historical high-impact weather information of wind farms, the periods when the weather has a strong impact on the wind farm can be selected for analysis to obtain the time points, frequency, and duration of high-impact weather events. The obtained data sources of different dimensions are added to the A data sources to obtain the A data sources, where A is a positive integer greater than 2.

[0038] Furthermore, as shown in Figure 2, step S300 of this embodiment further includes:

[0039] S310: Based on the historical high-impact weather information of the wind farm, extract the image set of the target wind farm during the historical high-impact weather of the wind farm from the image acquisition unit, perform real-time image brightness recognition analysis on the image set, and obtain a one-dimensional image data source;

[0040] S320: Based on the historical high-impact weather information of the wind farm, perform weather impact analysis on the weather of the target wind farm during historical periods to obtain a two-dimensional time period data source;

[0041] S330: Based on the historical high-impact weather information of the wind farm, the historical atmospheric sounding of the target wind farm is detected and analyzed to obtain a three-dimensional atmospheric sounding data source;

[0042] S340: Based on the historical high-impact weather information of the wind farm, perform numerical analysis on the historical weather data of the target wind farm to obtain a four-dimensional numerical data source;

[0043] S350: Add the one-dimensional image data source, the two-dimensional time period data source, the three-dimensional atmospheric sounding data source, and the four-dimensional numerical data source to the A data sources.

[0044] Specifically, based on the aforementioned historical high-impact weather information for wind farms, a collection of all image data from this historical information is extracted. By analyzing the brightness of real-time images, the ambient light intensity, cloud thickness, and other factors can be determined; this data constitutes a one-dimensional image data source. Based on the aforementioned historical high-impact weather information for wind farms, periods with strong weather impacts on the wind farms are selected for analysis. This allows for the determination of the time points, frequency, and duration of high-impact weather events occurring at the wind farms; this data constitutes a two-dimensional time period data source. Based on the aforementioned historical high-impact weather information for wind farms, the target wind farm is selected... Historical atmospheric sounding data, including ground meteorological observations, upper-air meteorological observations, atmospheric remote sensing, and meteorological satellite sounding, are analyzed to obtain atmospheric conditions and changes from the ground to the upper atmosphere, and from local to overall wind farm data. This data constitutes the three-dimensional atmospheric sounding data source. Based on the aforementioned historical high-impact weather information of wind farms, historical weather values ​​of the target wind farm are selected for analysis to obtain air pressure, humidity, temperature, dew point temperature, and other data within the wind farm. This data constitutes the four-dimensional numerical data source. All instances of the above four types of data sources together form A data sources, where A is a positive integer greater than 2.

[0045] S400: Based on the A data sources, perform feature pairing in the meteorological feature pairing library to determine B meteorological feature elements. Input the B meteorological feature elements and the A data sources into the meteorological pattern layer in the meteorological feature prediction model and output the meteorological feature information of the wind farm. Here, B is a positive integer greater than 1.

[0046] Specifically, the meteorological feature pairing library refers to a feature set formed by integrating all meteorological characteristics, such as air pressure, temperature, humidity, and wind speed, serving as a reference set for meteorological feature pairing. By pairing all instances of the aforementioned one-dimensional image data source, two-dimensional time period data source, three-dimensional atmospheric sounding data source, and four-dimensional numerical data source with the meteorological feature pairing library one by one, B meteorological feature elements can be obtained, where B is a positive integer greater than 1. The meteorological feature prediction model is a neural network model in machine learning that can continuously iterate and optimize itself. Inputting the B meteorological feature elements and the A data sources into the meteorological law layer of the meteorological feature prediction model outputs wind farm meteorological feature information, which can serve as reference data for subsequent multi-level meteorological factor response interval identification.

[0047] Furthermore, step S400 in this embodiment of the application also includes:

[0048] S410: The one-dimensional image data source is used to match image meteorological features in the meteorological feature matching library to obtain image brightness meteorological feature elements;

[0049] S420: The two-dimensional time period data source is paired with the time period meteorological features in the meteorological feature pairing database to obtain the time period-affected meteorological feature elements;

[0050] S430: The three-dimensional atmospheric sounding data source is paired with atmospheric state meteorological features in the meteorological feature pairing library to obtain atmospheric state meteorological feature elements;

[0051] S440: The four-dimensional numerical data source is paired with weather numerical meteorological features in the meteorological feature pairing library to obtain weather numerical meteorological feature elements;

[0052] S450: Add the image brightness meteorological feature element, the time period influence meteorological feature element, the atmospheric state meteorological feature element, and the weather value meteorological feature element to the B meteorological feature elements.

[0053] Specifically, the aforementioned one-dimensional image data source is paired one-to-one with the aforementioned meteorological feature pairing library. Image data sources with the same meteorological features are counted as one image brightness meteorological feature element, resulting in an image brightness meteorological feature element set. The aforementioned two-dimensional time period data source is paired one-to-one with the aforementioned meteorological feature pairing library. Time period data sources with the same meteorological features are counted as one time period data meteorological feature element, resulting in a time period impact meteorological feature element set. The aforementioned three-dimensional atmospheric sounding data source is paired one-to-one with the aforementioned meteorological feature pairing library. Atmospheric sounding data sources with the same meteorological features are counted as one atmospheric sounding meteorological feature element, resulting in an atmospheric state meteorological feature element set. The aforementioned four-dimensional numerical data source is paired one-to-one with the aforementioned meteorological feature pairing library. Four-dimensional numerical data sources with the same meteorological features are counted as one atmospheric sounding meteorological feature element, resulting in a weather numerical meteorological feature element set. The image brightness meteorological feature element set, the time period impact meteorological feature element set, the atmospheric state meteorological feature element set, and the weather numerical meteorological feature element set constitute B meteorological feature elements. The B meteorological feature elements can be used as the basic data for subsequent output of wind farm meteorological characteristic information.

[0054] Furthermore, step S400 in this embodiment of the application also includes:

[0055] S411: Assign a first weight to the image brightness meteorological feature elements;

[0056] S421: Assign a second weight to the meteorological characteristic elements affecting the time period;

[0057] S431: Assign a third weight to the atmospheric state meteorological characteristic elements;

[0058] S441: Assign a fourth weight to the meteorological feature elements of the weather data;

[0059] S451: Integrate the first weight, the second weight, the third weight, and the fourth weight, and update the B meteorological feature elements according to different weight ratios.

[0060] Specifically, based on the accuracy of the four types of feature elements in weather prediction, the image brightness meteorological feature elements, time period influence meteorological feature elements, atmospheric state meteorological feature elements, and weather value meteorological feature elements are weighted accordingly. The weight ratio of each feature element is proportional to its accuracy in weather prediction. For example, based on the accuracy of the image brightness meteorological feature element in weather prediction, its weight allocation coefficient is 1, and the corresponding weight ratio of the four types of feature elements is 1:2:4:3. A first weight is assigned to the image brightness meteorological feature element, a second weight to the time period influence meteorological feature element, a third weight to the atmospheric state meteorological feature element, and a fourth weight to the weather value meteorological feature element. The first weight, the second weight, the third weight, and the fourth weight are integrated, and the B meteorological feature elements are updated according to different weight ratios, which can improve the accuracy of the B meteorological feature elements and thus improve the accuracy of the output wind farm meteorological characteristic information.

[0061] Furthermore, as shown in Figure 3, step S400 of this embodiment further includes:

[0062] S460: Obtain meteorological source characteristic data of the target wind farm based on the B weighted meteorological feature elements and the A data sources;

[0063] S470: Based on the meteorological source characteristic data of the target wind farm, obtain air pressure source data, humidity source data, temperature source data, and wind speed source data;

[0064] S480: Input the air pressure source data, the humidity source data, the temperature source data, and the wind speed source data into the meteorological law layer of the meteorological characteristic prediction model to generate air pressure law feature data, humidity law feature data, temperature law feature data, and wind speed law feature data;

[0065] S490: Normalize the air pressure pattern characteristic data, the humidity pattern characteristic data, the temperature pattern characteristic data, and the wind speed pattern characteristic data, and output the meteorological characteristic information of the wind farm.

[0066] Specifically, the meteorological source feature data of the target wind farm refers to the meteorological features contained in each data source, including air pressure, temperature, humidity, etc.; the meteorological source feature data of the target wind farm is extracted from the B weighted meteorological feature elements and the A data sources, and then air pressure source data, humidity source data, temperature source data, and wind speed source data are obtained from the meteorological source feature data of the target wind farm.

[0067] The meteorological feature prediction model is a neural network model in machine learning that can continuously iterate and optimize itself. It can be obtained using training and supervised datasets. Each training dataset includes data from air pressure, humidity, temperature, and wind speed sources. The supervised dataset contains meteorological data corresponding to the training dataset. The construction process involves first inputting each training dataset into the model, then adjusting the model's output using the corresponding supervised data. If the model's output matches the supervised data, training for that group ends. This process is repeated with all data in the training dataset until training is complete. To ensure accuracy, the model can be tested using a test dataset with an accuracy rate of 85%. If the accuracy rate on the test dataset meets 85%, the meteorological feature prediction model is considered successfully constructed.

[0068] Finally, the air pressure source data, humidity source data, temperature source data, and wind speed source data are input into the meteorological law layer of the meteorological characteristic prediction model, and the air pressure law feature data, humidity law feature data, temperature law feature data, and wind speed law feature data are output.

[0069] Specifically, normalization uses mathematical functions to map all data to the same scale. Because the obtained data varies greatly in size, there are outlier samples. These outlier samples involve large number calculations, which are very time-consuming and produce abnormally large results. Furthermore, the weighting distribution is uneven, with larger numbers potentially receiving greater weight. However, since large numbers are not necessarily the most critical factor determining the data outcome, this can lead to inaccurate predictions. The purpose of normalization is to confine the preprocessed data to a certain range, such as [0,1] or [-1,1], thereby eliminating the adverse effects of outlier samples. Normalizing the air pressure, humidity, temperature, and wind speed characteristic data outputs wind farm meteorological characteristic information, which can serve as the basic data for identifying multi-level meteorological factor response intervals.

[0070] S500: Based on the historical wind farm meteorological data and the wind farm meteorological characteristic information, identify the response intervals of multi-level meteorological factors, and generate multi-level meteorological factor sensitivity indices based on the identification results;

[0071] Specifically, the aforementioned historical meteorological data and meteorological characteristics of wind farms are used as basic data to identify multi-level meteorological factor response intervals. These multi-level meteorological factor response intervals refer to emergency response intervals divided according to the degree of meteorological impact, development trend, and potential hazards. By identifying these multi-level meteorological factor response intervals, emergency response intervals can be divided for the historical meteorological data of the target wind farm. The identification results can generate a multi-level meteorological factor sensitivity index, which can serve as a reference standard for subsequent meteorological forecasts.

[0072] Furthermore, step S500 in this embodiment of the application also includes:

[0073] S510: Collect historical wind and rain information of wind farms based on the historical meteorological data of wind farms, and obtain the historical wind and rain information collection results;

[0074] S520: Obtain the meteorological impact level data of wind and rain based on the analysis of the historical wind and rain information collection results;

[0075] S530: Adjust the meteorological characteristics information of the wind farm based on the wind and rain meteorological impact level data to obtain the wind and rain meteorological characteristics adjustment results;

[0076] S540: Identify the multi-level meteorological factor response intervals based on the adjustment results of the wind and rain meteorological characteristics, and generate a multi-level meteorological factor sensitivity index based on the identification results.

[0077] Specifically, the historical wind farm wind and rain information refers to the data information on wind and rain weather in the target wind farm over a period of time. Collecting historical wind and rain information based on the aforementioned historical wind farm meteorological data involves filtering out meteorological data related to wind and rain events to form historical wind and rain information collection results. Analyzing these historical wind and rain information results involves classifying all wind and rain information according to wind force level, rainfall, duration, and the degree of impact on the wind farm's operation to obtain wind and rain meteorological impact level data. Adjusting the wind farm meteorological characteristic information based on the wind and rain meteorological impact level data involves classifying and sorting the meteorological characteristic information according to the wind and rain meteorological impact level data to obtain wind and rain meteorological characteristic adjustment results.

[0078] Specifically, multi-level meteorological factor response intervals refer to emergency response intervals divided according to the degree of meteorological impact, development trend, and potential hazards. For example, rainstorm warning signals are divided into four levels, represented by blue, yellow, orange, and red, with each level representing a rainfall interval. Blue rainstorm warning signal: Rainfall will reach or has already reached 50 mm within 12 hours, potentially causing or already causing impact, and the rainfall is likely to continue; Yellow rainstorm warning signal: Rainfall will reach or has already reached 50 mm within 6 hours, potentially causing or already causing impact, and the rainfall is likely to continue; Orange rainstorm warning signal: Rainfall will reach or has already reached 50 mm within 3 hours, potentially causing or already causing significant impact, and the rainfall is likely to continue; Red rainstorm warning signal: Rainfall will reach or has already reached 100 mm within 3 hours, potentially causing or already causing severe impact, and the rainfall is likely to continue. The adjustment results of the aforementioned wind and rain meteorological characteristics are identified one by one with the meteorological factor response intervals to determine which interval each meteorological characteristic falls within. Finally, a multi-level meteorological factor sensitivity index is generated based on the identification results. For example, rainstorm warning signals are divided into four levels according to rainfall, duration, and impact. The multi-level meteorological factor sensitivity index can also be classified according to the magnitude of each influencing factor from high to low. The multi-level meteorological factor sensitivity index can serve as a reference standard for subsequent weather forecasts.

[0079] The S600 collects meteorological factor prediction data and forecasts the weather for the target wind farm based on the meteorological factor prediction data and the multi-level meteorological factor sensitivity index.

[0080] Specifically, current meteorological data of the target wind farm is collected, and the interaction between various meteorological factors is considered. The data is then processed to generate meteorological factor prediction data. By combining the meteorological factor prediction data with the aforementioned multi-level meteorological factor sensitivity index, the weather forecast for the target wind farm can be made, which can improve the accuracy of the prediction data.

[0081] Furthermore, step S600 in this embodiment of the application also includes:

[0082] S610: Obtains wind data, rainfall data, and temperature data of the target wind farm through the sensing unit;

[0083] S620: Perform a linear correlation between the wind data, the rainfall data, and the temperature data to obtain a linear correlation result. The linear correlation formula is as follows:

[0084]

[0085] Where D(X)>0, D(Y)>0, D(Z)>0, X represents wind data, Y represents rain data, and Z represents temperature data.

[0086] S630: Based on the linear correlation results, generate meteorological factor prediction data.

[0087] Specifically, the sensing unit refers to meteorological monitoring equipment that can continuously monitor environmental data such as ambient temperature, relative humidity, wind speed, wind direction, atmospheric pressure, rainfall, photoelectric radiation, PM2.5, and PM10 24 hours a day, and transmit the analyzed and processed data to the public in one go. The sensing unit obtains wind, rainfall, and temperature data for the target wind farm. These data, including wind speed, wind direction, heavy rain, high temperature, and low temperature, affect the meteorological conditions of the wind farm. The wind, rainfall, and temperature data are linearly correlated; wind speed affects rain direction, and temperature affects whether rain turns to snow. Linear correlation means that there is a functional relationship between several variables, and a change in one variable affects the others. The linear correlation result, i.e., the functional relationship between the wind, rainfall, and temperature data, is expressed by the following formula:

[0088]

[0089] Where D(X)>0, D(Y)>0, D(Z)>0, X represents wind data, Y represents rainfall data, and Z represents temperature data. Based on the linear correlation formula, adjusting the wind, rainfall, and temperature data yields more accurate meteorological factor prediction data.

[0090] In summary, the embodiments of this application have at least the following technical effects:

[0091] This application acquires historical wind farm meteorological data by connecting to a meteorological platform; then analyzes the historical wind farm meteorological data to obtain historical high-impact weather information; further, it performs multi-dimensional data source analysis on the historical high-impact weather information to determine A data sources; based on the A data sources, it performs feature pairing in a meteorological feature pairing library to determine B meteorological feature elements; then, it inputs the B meteorological feature elements and the A data sources into the meteorological law layer of a meteorological characteristic prediction model to obtain wind farm meteorological characteristic information; then, based on the historical wind farm meteorological data and the wind farm meteorological characteristic information, it identifies multi-level meteorological factor response intervals and generates multi-level meteorological factor sensitivity indices from the identification results; finally, it collects meteorological factor prediction data and forecasts the weather of the target wind farm based on the meteorological factor prediction data and the multi-level meteorological factor sensitivity indices.

[0092] This technology enables multi-source data-driven meteorological forecasting of wind farms, thereby improving the accuracy of meteorological forecast results.

[0093] Example 2

[0094] Based on the same inventive concept as the wind farm key meteorological factor forecasting method based on multi-source data in the foregoing embodiments, as shown in Figure 4, this application provides a wind farm key meteorological factor forecasting system based on multi-source data. The system and method embodiments in this application are based on the same inventive concept. The system includes:

[0095] Meteorological data acquisition module 11, which is used to acquire historical wind farm meteorological data by connecting to a meteorological platform;

[0096] Data analysis module 12 is used to analyze the historical meteorological data of the wind farm to obtain historical high-impact weather information of the wind farm;

[0097] Multidimensional data source analysis module 13 is used to perform multidimensional data source analysis on the historical high-impact weather information of wind farms to determine A data sources;

[0098] The meteorological characteristic information acquisition module 14 is used to perform feature matching of the meteorological characteristic matching library based on the A data sources, determine B meteorological characteristic elements, input the B meteorological characteristic elements and the A data sources into the meteorological law layer in the meteorological characteristic prediction model, and output the meteorological characteristic information of the wind farm, where B is a positive integer greater than 1.

[0099] The multi-level meteorological factor sensitivity index generation module 15 is used to identify the response range of multi-level meteorological factors based on the historical wind farm meteorological data and the wind farm meteorological characteristic information, and generate a multi-level meteorological factor sensitivity index based on the identification results.

[0100] The weather forecast module 16 is used to collect meteorological factor prediction data and forecast the weather of the target wind farm based on the meteorological factor prediction data and the multi-level meteorological factor sensitivity index.

[0101] Furthermore, the system also includes:

[0102] The image brightness recognition and analysis module is used to extract the image set of the target wind farm during the historical high-impact weather information of the wind farm based on the historical high-impact weather information of the wind farm, and to perform real-time image brightness recognition and analysis on the image set to obtain a one-dimensional image data source.

[0103] The weather impact analysis module is used to perform weather impact analysis on the weather of the target wind farm during historical periods based on the historical high-impact weather information of the wind farm, and obtain a two-dimensional time period data source;

[0104] An atmospheric sounding and analysis module is used to detect and analyze the historical atmospheric sounding of the target wind farm based on the historical high-impact weather information of the wind farm, and obtain a three-dimensional atmospheric sounding data source.

[0105] The numerical analysis module is used to perform numerical analysis on the historical weather data of the target wind farm based on the historical high-impact weather information of the wind farm, and obtain a four-dimensional numerical data source.

[0106] Furthermore, the system also includes:

[0107] An image brightness meteorological feature pairing module is used to pair the image meteorological features of the one-dimensional image data source in the meteorological feature pairing library to obtain image brightness meteorological feature elements.

[0108] A time-period meteorological feature pairing module is used to pair the two-dimensional time-period data source with time-period meteorological features in the meteorological feature pairing database to obtain time-period meteorological feature elements.

[0109] An atmospheric state meteorological feature pairing module is used to pair the atmospheric state meteorological features of the three-dimensional atmospheric sounding data source in the meteorological feature pairing library to obtain atmospheric state meteorological feature elements.

[0110] A weather data meteorological feature pairing module is used to pair the four-dimensional numerical data source with weather data meteorological features in the meteorological feature pairing library to obtain weather data meteorological feature elements.

[0111] Furthermore, the system also includes:

[0112] The weight allocation module is used to allocate a first weight to the image brightness meteorological feature element; allocate a second weight to the time period influence meteorological feature element; allocate a third weight to the atmospheric state meteorological feature element; and allocate a fourth weight to the weather value meteorological feature element.

[0113] Furthermore, the system also includes:

[0114] A meteorological source feature data acquisition module is used to obtain meteorological source feature data of a target wind farm based on the B weighted meteorological feature elements and the A data sources.

[0115] The meteorological characteristic prediction model module is used to input the air pressure source data, the humidity source data, the temperature source data, and the wind speed source data into the meteorological law layer of the meteorological characteristic prediction model to generate air pressure law feature data, humidity law feature data, temperature law feature data, and wind speed law feature data.

[0116] The normalization processing module is used to normalize the air pressure pattern feature data, the humidity pattern feature data, the temperature pattern feature data, and the wind speed pattern feature data, and output the meteorological characteristic information of the wind farm.

[0117] Furthermore, the system also includes:

[0118] The historical wind and rain information acquisition module is used to collect historical wind and rain information of the wind farm based on the historical meteorological data of the wind farm, and obtain the historical wind and rain information acquisition results.

[0119] The result analysis module is used to obtain wind and rain meteorological impact level data based on the historical wind and rain information collection results.

[0120] A meteorological characteristic adjustment module is used to adjust the meteorological characteristic information of the wind farm according to the wind and rain meteorological impact level data to obtain the wind and rain meteorological characteristic adjustment result;

[0121] The meteorological factor response interval identification module is used to identify multi-level meteorological factor response intervals based on the adjustment results of the wind and rain meteorological characteristics, and generate multi-level meteorological factor sensitivity indices based on the identification results.

[0122] Furthermore, the system also includes:

[0123] A wind farm data acquisition module is used to obtain wind data, rainfall data, and temperature data of the target wind farm through a sensing unit.

[0124] The linear correlation module is used to linearly correlate the wind data, the rain data, and the temperature data to obtain a linear correlation result. The linear correlation formula is as follows:

[0125]

[0126] Where D(X)>0, D(Y)>0, D(Z)>0, X represents wind data, Y represents rain data, and Z represents temperature data;

[0127] A meteorological factor prediction data generation module generates meteorological factor prediction data based on the linear correlation results.

[0128] 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, specific embodiments have been described above. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps described in the claims can be performed in a different order than that shown in the embodiments and still achieve the desired result. Additionally, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0129] 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.

[0130] 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 modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A method for forecasting key meteorological factors of wind farms based on multi-source data, characterized in that, This method is applied to a wind farm key meteorological factor forecasting system based on multi-source data. The system includes an image acquisition unit and a sensing unit. The method includes: acquiring historical wind farm meteorological factor data by connecting to a meteorological platform; analyzing the historical wind farm meteorological factor data to obtain historical high-impact weather information; performing multi-dimensional data source analysis on the historical high-impact weather information to determine A data sources, where A is a positive integer greater than 2; performing feature pairing in a meteorological feature pairing library based on the A data sources to determine B meteorological feature elements; inputting the B meteorological feature elements and the A data sources into the meteorological law layer of a meteorological characteristic prediction model to output wind farm meteorological characteristic information, where B is a positive integer greater than 1; identifying multi-level meteorological factor response intervals based on the historical wind farm meteorological data and the wind farm meteorological characteristic information; generating multi-level meteorological factor sensitivity indices based on the identification results; and collecting meteorological factor prediction data. Based on the meteorological factor prediction data and the multi-level meteorological factor sensitivity index, key meteorological factors of the target wind farm are forecasted; A data sources are determined. The method further includes: based on the historical high-impact weather information of the wind farm, extracting the image set of the target wind farm during historical high-impact weather, performing real-time image brightness recognition analysis on the image set to obtain a one-dimensional image data source; based on the historical high-impact weather information of the wind farm, performing weather impact analysis on the weather of the target wind farm in historical periods to obtain a two-dimensional time period data source; based on the historical high-impact weather information of the wind farm, performing detection analysis on the historical atmospheric sounding of the target wind farm to obtain a three-dimensional atmospheric sounding data source; based on the historical high-impact weather information of the wind farm, performing numerical analysis on the historical weather values ​​of the target wind farm to obtain a four-dimensional numerical data source; and adding the one-dimensional image data source, the two-dimensional time period data source, the three-dimensional atmospheric sounding data source, and the four-dimensional numerical data source to the A data sources.

2. The method as described in claim 1, characterized in that, The method for determining B meteorological feature elements further includes: pairing image meteorological features of a one-dimensional image data source with those of the meteorological feature pairing library to obtain image brightness meteorological feature elements; pairing time period meteorological features of a two-dimensional time period data source with those of the meteorological feature pairing library to obtain time period impact meteorological feature elements; pairing atmospheric state meteorological features of a three-dimensional atmospheric sounding data source with those of the meteorological feature pairing library to obtain atmospheric state meteorological feature elements; pairing weather numerical meteorological features of a four-dimensional numerical data source with those of the meteorological feature pairing library to obtain weather numerical meteorological feature elements; and adding the image brightness meteorological feature elements, the time period impact meteorological feature elements, the atmospheric state meteorological feature elements, and the weather numerical meteorological feature elements to the B meteorological feature elements.

3. The method as described in claim 2, characterized in that, The method further includes determining B meteorological feature elements: assigning a first weight to the image brightness meteorological feature element; assigning a second weight to the time period influence meteorological feature element; assigning a third weight to the atmospheric state meteorological feature element; assigning a fourth weight to the weather value meteorological feature element; integrating the first weight, the second weight, the third weight, and the fourth weight, and updating the B meteorological feature elements according to different weight ratios.

4. The method as described in claim 1, outputting wind farm meteorological characteristic information, characterized in that, The method further includes: obtaining meteorological source feature data of the target wind farm based on B weighted meteorological feature elements and A data sources; obtaining air pressure source data, humidity source data, temperature source data, and wind speed source data based on the meteorological source feature data of the target wind farm; inputting the air pressure source data, humidity source data, temperature source data, and wind speed source data into the meteorological law layer of the meteorological characteristic prediction model to generate air pressure law feature data, humidity law feature data, temperature law feature data, and wind speed law feature data; and normalizing the air pressure law feature data, humidity law feature data, temperature law feature data, and wind speed law feature data to output wind farm meteorological characteristic information.

5. The method as described in claim 1, characterized in that, The method for identifying multi-level meteorological factor response intervals further includes: collecting historical wind and rain information from the historical wind farm meteorological data to obtain historical wind and rain information collection results; analyzing the historical wind and rain information collection results to obtain wind and rain meteorological impact level data; adjusting the wind farm meteorological characteristic information according to the wind and rain meteorological impact level data to obtain wind and rain meteorological characteristic adjustment results; identifying multi-level meteorological factor response intervals from the wind and rain meteorological characteristic adjustment results; and generating multi-level meteorological factor sensitivity indices based on the identification results.

6. The method as described in claim 1, characterized in that, The method for collecting meteorological factor prediction data further includes: obtaining wind data, rainfall data, and temperature data of the target wind farm through a sensing unit; and performing a linear correlation between the wind data, rainfall data, and temperature data to obtain a linear correlation result, wherein the linear correlation formula is as follows: in, X represents wind data, Y represents rainfall data, and Z represents temperature data; S630 generates meteorological factor prediction data based on the linear correlation results.

7. A wind farm key meteorological factor forecasting system based on multi-source data, characterized in that, The system includes: a meteorological data acquisition module, which acquires historical wind farm meteorological factor data by connecting to a meteorological platform; a data analysis module, which analyzes the historical wind farm meteorological factor data to obtain historical high-impact weather information; a multi-dimensional data source analysis module, which performs multi-dimensional data source analysis on the historical high-impact weather information to determine A data sources; a meteorological feature information acquisition module, which performs feature matching in a meteorological feature matching library based on the A data sources to determine B meteorological feature elements, inputs the B meteorological feature elements and the A data sources into the meteorological law layer of a meteorological feature prediction model, and outputs wind farm meteorological feature information, where B is a positive integer greater than 1; a multi-level meteorological factor sensitivity index generation module, which identifies multi-level meteorological factor response intervals based on the historical wind farm meteorological data and the wind farm meteorological feature information, and generates multi-level meteorological factor sensitivity indices based on the identification results; and a meteorological forecasting module. The system comprises the following modules: a meteorological forecasting module for collecting meteorological factor prediction data and forecasting key meteorological factors of the target wind farm based on the meteorological factor prediction data and the multi-level meteorological factor sensitivity index; an image brightness recognition and analysis module for extracting image sets of the target wind farm during historical high-impact weather conditions based on historical wind farm high-impact weather information, performing real-time image brightness recognition and analysis on the image sets to obtain a one-dimensional image data source; a weather impact analysis module for performing weather impact analysis on the target wind farm during historical periods based on historical high-impact weather information to obtain a two-dimensional period data source; an atmospheric sounding analysis module for performing historical atmospheric sounding analysis on the target wind farm based on historical high-impact weather information to obtain a three-dimensional atmospheric sounding data source; and a numerical analysis module for performing numerical analysis on historical weather values ​​of the target wind farm based on historical high-impact weather information to obtain a four-dimensional numerical data source.

Images