Method and device for predicting new energy output

Abstract

The present disclosure provides a prediction method, device, electronic device, computer-readable storage medium and computer program product for new energy output, which belongs to the field of new energy power generation technology. The prediction method includes: obtaining at least two meteorological data from different sources, processing the at least two meteorological data in time and space to obtain processed meteorological data; extracting features from the processed meteorological data to obtain complexity feature information and dynamic feature information of meteorological conditions corresponding to the at least two meteorological data; constructing a random evolution model of meteorological conditions based on the complexity feature information and dynamic feature information of meteorological conditions; sampling the random evolution model at least twice to generate at least two different meteorological scenes; and predicting new energy output based on at least two different meteorological scenes. The prediction method disclosed in the present disclosure can simultaneously strike a balance between prediction accuracy, scene diversity and computational complexity.

Description

Technical Field

[0001] The present invention belongs to the technical field of renewable energy power generation, and specifically relates to a method, device, electronic equipment, computer-readable storage medium and computer program product for predicting renewable energy output. Background Art

[0002] As the global demand for renewable energy continues to grow, the proportion of new energy (such as wind and solar energy) in the power system has gradually increased. However, renewable energy power generation is significantly affected by meteorological conditions. Changes in meteorological conditions often lead to increased uncertainty in renewable energy output, posing challenges to the stable operation of the power system. In order to effectively address this issue, the planning and scheduling of power systems need to rely on accurate prediction of renewable energy output and multi-scenario generation. However, existing renewable energy output prediction methods have significant deficiencies when dealing with complex and changeable meteorological conditions. Summary of the invention

[0003] The present invention aims to solve at least one of the technical problems existing in the prior art and to provide a prediction method, a prediction device, a computer-readable storage medium and a computer program product for new energy output that can simultaneously strike a balance between prediction accuracy, scenario diversity and computational complexity.

[0004] In order to achieve the above-mentioned purpose, in a first aspect, an embodiment of the present disclosure provides a method for predicting new energy output, comprising: obtaining at least two types of meteorological data from different sources, wherein the meteorological data at least includes two of ground meteorological station data, satellite remote sensing data, and meteorological radar data;

[0005] Performing temporal and spatial processing on the at least two types of meteorological data to obtain processed meteorological data;

[0006] Performing feature extraction on the processed meteorological data to obtain complexity feature information and dynamic feature information of meteorological conditions corresponding to the multiple meteorological data;

[0007] Constructing a random evolution model of the meteorological conditions according to the complexity characteristic information and dynamic characteristic information of the meteorological conditions;

[0008] Sampling the random evolution model at least twice to generate at least two different meteorological scenarios;

[0009] Based on at least two different meteorological scenarios, the output of the renewable energy is predicted.

[0010] In some embodiments, the ground weather station data includes at least one of the following: ground temperature data, humidity data, wind speed data, and air pressure data;

[0011] The satellite remote sensing data includes at least one of the following: solar radiation data, cloud cover data, and surface temperature data;

[0012] The weather radar data includes precipitation intensity data and / or wind speed distribution data.

[0013] In some embodiments, performing temporal and spatial processing on the at least two types of meteorological data to obtain processed meteorological data includes:

[0014] Unifying the at least two meteorological data to the same time resolution, and filling in missing data in the at least two meteorological data due to asynchronous data collection time, to obtain time-aligned meteorological data;

[0015] The time-aligned meteorological data are unified into the same spatial resolution, and missing data in the time-aligned meteorological data due to different spatial resolutions of data are filled to obtain the processed meteorological data.

[0016] In some embodiments, the feature extraction of the processed meteorological data to obtain the complexity feature information of the meteorological conditions corresponding to the at least two meteorological data includes:

[0017] Dividing the time series of the processed meteorological data into different scales, and calculating the minimum number of boxes required to cover all the processed meteorological data at each scale;

[0018] In a double logarithmic coordinate system, a complexity characteristic curve is fitted according to different scales of division and the minimum number of boxes;

[0019] According to the complexity characteristic curve, complexity characteristic information of the meteorological conditions corresponding to the at least two meteorological data is obtained.

[0020] In some embodiments, the extracting features of the processed meteorological data to obtain dynamic feature information of meteorological conditions corresponding to the at least two meteorological data includes:

[0021] Mapping the one-dimensional time series in the time series of the processed meteorological data into a high-dimensional phase space;

[0022] The exponential separation rate of adjacent points in the high-dimensional phase space is calculated, and the dynamic characteristic information of the meteorological conditions corresponding to the at least two meteorological data is determined according to the exponential separation rate.

[0023] In some embodiments, constructing the random evolution model of the meteorological condition according to the complexity characteristic information and dynamic characteristic information of the meteorological condition includes:

[0024] According to the complexity characteristic information and dynamic characteristic information of the meteorological conditions, statistical analysis is performed on the time series of the at least two meteorological data to determine the parameters of the drift term and the expansion term;

[0025] Taking meteorological variables as variables of the random evolution model, and based on the parameters of the drift term and the extension term, an initial random evolution model is obtained, wherein the drift term represents the average change trend of the meteorological variable, and the extension term represents the random fluctuation of the meteorological variable;

[0026] The initial random evolution model is discretized to obtain the random evolution model to simulate the future evolution path of the meteorological conditions.

[0027] In some embodiments, the sampling of the random evolution model at least twice to generate at least two different meteorological scenarios includes:

[0028] Selecting an initial meteorological state from the at least two meteorological data, and generating a candidate state according to the initial meteorological state;

[0029] By comparing the posterior probability of the current state with the candidate state, calculating the acceptance probability of the candidate state;

[0030] Determine the current final state based on the acceptance probability and the randomly generated random number, and use the current final state as a sampling result;

[0031] At least two different meteorological scenarios are generated based on at least two sampling results.

[0032] In some embodiments, the forecasting of the new energy output based on at least two different meteorological scenarios includes:

[0033] Performing at least two objective optimizations on at least two different meteorological scenarios to obtain an optimized meteorological scenario;

[0034] The output of the renewable energy is predicted based on the optimized meteorological scenario.

[0035] In some embodiments, the at least two different meteorological scenarios are subjected to at least two objective optimizations to obtain an optimized meteorological scenario, including:

[0036] Based on at least two different meteorological scenarios, at least two optimization objectives are determined, and at least two objective optimization functions are constructed according to the at least two optimization objectives; wherein the optimization objectives include prediction accuracy, scenario diversity, and computational complexity;

[0037] Optimizing the at least two objective optimization functions to obtain an optimization result;

[0038] Determine a non-inferior solution in which the at least two optimization objectives in the optimization result are balanced, and select the optimal solution from the non-inferior solutions to ensure that the generated meteorological scenario achieves the best balance between prediction accuracy, scenario diversity and computational complexity, so as to obtain an optimized meteorological scenario.

[0039] In some embodiments, the forecasting of the new energy output based on at least two different meteorological scenarios includes:

[0040] Extract key meteorological parameters from at least two different meteorological scenarios, input them into a preset new energy output prediction model, and calculate the power output of wind power generation and the power output of photovoltaic power generation, wherein the key meteorological data include wind speed data and solar radiation intensity data;

[0041] Based on the power output of the wind power generation and the power output of the photovoltaic power generation, a total new energy output scenario at different time points is obtained;

[0042] The new energy output is predicted based on the total new energy output scenario.

[0043] In a second aspect, the present disclosure further provides a device for calculating the short-circuit ratio of a new energy device connected to the grid, comprising:

[0044] A data acquisition module, used to obtain a variety of meteorological data from different sources, wherein the multiple meteorological data at least include ground meteorological station data, satellite remote sensing data and meteorological radar data;

[0045] A data processing module, used for performing temporal and spatial processing on the various meteorological data to obtain processed meteorological data;

[0046] A feature extraction module, used to extract features from the processed meteorological data to obtain complexity feature information and dynamic feature information of meteorological conditions corresponding to the various meteorological data;

[0047] A stochastic modeling module, used for constructing a stochastic evolution model of the meteorological conditions according to the complexity characteristic information and dynamic characteristic information of the meteorological conditions;

[0048] A scenario generation module, used to perform multiple sampling on the random evolution model to generate multiple different meteorological scenarios;

[0049] The prediction module is used to predict the output of the renewable energy based on a plurality of different meteorological scenarios.

[0050] In some embodiments, the data processing module includes:

[0051] A time processing module, used to unify the at least two meteorological data to the same time resolution, and fill in the missing data in the at least two meteorological data due to the asynchronization of data collection time, so as to obtain the time-aligned meteorological data;

[0052] The spatial processing module is used to unify the time-aligned meteorological data into the same spatial resolution and fill in the missing data in the time-aligned meteorological data due to spatial asynchrony of data collection to obtain the processed meteorological data.

[0053] In some embodiments, the prediction module includes:

[0054] A calculation module, used to extract key meteorological parameters in at least two different meteorological scenes, input them into a preset new energy output prediction model, and calculate the power output of wind power generation and photovoltaic power generation, wherein the key meteorological data include wind speed data and solar radiation intensity data;

[0055] An output generation module, used to generate total renewable energy output scenarios at different time points based on the power output of the wind power generation and the power output of the photovoltaic power generation;

[0056] A prediction submodule is used to predict the new energy output based on the total new energy output scenario.

[0057] In some embodiments, the prediction module further includes:

[0058] An optimization module is used to perform multi-objective optimization on multiple different meteorological scenarios to obtain optimized meteorological scenarios;

[0059] The calculation module is specifically used to extract key meteorological parameters in the optimized meteorological scene, input them into a preset new energy output prediction model, and calculate the power output of wind power generation and the power output of photovoltaic power generation.

[0060] In a third aspect, the present disclosure further provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the computing method of the first aspect described above is implemented.

[0061] In a fourth aspect, the present disclosure further provides a computer-readable storage medium storing a computer program for executing the computing method of the first aspect.

[0062] In a fifth aspect, the present disclosure further provides a computer program product, including a computer program / instruction, which implements the steps of the computing method of the first aspect when the computer program / instruction is executed by a processor.

[0063] The present disclosure can effectively integrate various types of meteorological information to form a high-quality, widely covered data set by fusing multi-source data from ground meteorological stations, satellite remote sensing and meteorological radar. The fusion of multi-source data overcomes the limitations of a single data source, making the description of meteorological conditions more comprehensive and accurate, and providing a solid data foundation for subsequent modeling and scene generation. Further, the embodiment of the present disclosure solves the problem of resolution differences in time and space of multi-source meteorological data by performing time processing and space processing on various types of meteorological data, ensuring the consistency of data on a unified space-time grid. The data processed in this way not only improves the accuracy of model input, but also avoids errors caused by data inconsistency, ensuring the reliability of subsequent feature extraction and modeling. Further, the present disclosure extracts the complexity and nonlinear dynamic behavior of meteorological conditions, constructs a random evolution model, captures the random volatility of meteorological conditions, and then generates meteorological scenes that can cover a wide range of changes in meteorological conditions based on meteorological conditions, providing diversified inputs for new energy output prediction, enhancing the system's ability to cope with uncertainty, and making the final prediction results more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] In order to more clearly illustrate the embodiments of the present disclosure or the technical solutions in the prior art, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0065] Figure 1 A flowchart of a method for predicting new energy output provided by an embodiment of the present disclosure;

[0066] Figure 2 A flowchart of a specific execution process of step S3 provided in an embodiment of the present disclosure;

[0067] Figure 3 A flowchart for executing step S4 provided in an embodiment of the present disclosure;

[0068] Figure 4 A structural block diagram of a new energy output prediction device provided in an embodiment of the present disclosure;

[0069] Figure 5 A structural block diagram of a feature extraction module provided in an embodiment of the present disclosure;

[0070] Figure 6 A structural block diagram of a prediction module provided by an embodiment of the present disclosure;

[0071] Figure 7 A structural block diagram of an electronic device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0072] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention is further described in detail and completely in conjunction with the accompanying drawings and specific implementation methods. Obviously, the described embodiments are only part of the embodiments of the present disclosure, rather than all the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present disclosure.

[0073] Traditional renewable energy output forecasting methods usually rely on meteorological data from a single source, which limits the comprehensive understanding of meteorological conditions. In addition, existing methods often use simple linear interpolation or statistical regression models in meteorological data processing, which have limited prediction accuracy when faced with the nonlinear characteristics of meteorological systems.

[0074] In the modeling process, existing methods usually ignore the randomness and complex dynamic behavior of meteorological conditions and lack an effective description of the uncertainties in meteorological conditions. Even though some methods try to cover the changes in meteorological conditions through scenario generation, these generation methods often fail to strike a balance between prediction accuracy, scenario diversity, and computational complexity, resulting in the lack of representativeness of the generated scenarios or excessive consumption of computing resources.

[0075] In response to the above-mentioned problems existing in the prior art, the present disclosure provides a new energy output scenario generation solution based on historical meteorological characteristics, which generates new energy output scenarios with high precision, diversity and low computational complexity under changeable and complex meteorological conditions to support the stable operation and optimized scheduling of the power system.

[0076] In a first aspect, an embodiment of the present disclosure provides a method for predicting new energy output.

[0077] Figure 1 A flowchart of a method for predicting new energy output provided by an embodiment of the present disclosure. Figure 1 As shown, the method includes:

[0078] S1. Acquire at least two types of meteorological data from different sources, wherein the meteorological data includes at least two of ground meteorological station data, satellite remote sensing data and meteorological radar data.

[0079] Specifically, due to the differences in temporal and spatial resolutions of different meteorological data sources, a single data source is often difficult to fully and accurately capture the dynamic changes in meteorological conditions. The present disclosure integrates meteorological data from a variety of different sources to provide basic data support for subsequent modeling and generation of new energy output scenarios. These data sources include at least two of ground meteorological stations, satellite remote sensing systems, and meteorological radars, which provide information on various meteorological elements at different temporal and spatial resolutions.

[0080] In some embodiments, the ground weather station data includes at least one of the following: ground temperature data, humidity data, wind speed data, and air pressure data. The ground weather station data usually has a time resolution of 10 minutes to 1 hour, providing high-precision ground observation data suitable for local meteorological monitoring.

[0081] In some embodiments, the satellite remote sensing data includes at least one of the following: solar radiation data, cloud cover data, and surface temperature. The temporal resolution of the satellite remote sensing data is usually 15 minutes to 1 hour, and the spatial resolution can vary depending on the type of satellite. The advantage of satellite data is its large coverage and high spatial resolution, which is suitable for monitoring macroscopic phenomena such as clouds and radiation in the atmosphere.

[0082] In some embodiments, the weather radar data includes precipitation intensity data and / or wind speed distribution data. Among them, the weather radar data is mainly used to obtain high-resolution meteorological information such as precipitation intensity and wind speed distribution. Weather radar usually provides a time resolution of several minutes and can accurately capture small-scale convective weather phenomena, such as thunderstorms and local heavy precipitation events.

[0083] In the disclosed embodiments, by acquiring meteorological data from different sources, information on various meteorological elements can be provided at different time and spatial resolutions, providing basic data support for subsequent modeling and generation of new energy output scenarios.

[0084] S2. Process the at least two types of meteorological data in time and space to obtain processed meteorological data.

[0085] Specifically, after obtaining meteorological data from different sources, these data need to be further processed to ensure their consistency in time and space. Since data from different sources have different temporal and spatial resolutions, by performing temporal and spatial processing on the data, the meteorological information from various data sources can be integrated into a unified spatiotemporal grid, providing consistent and complete input data for subsequent feature extraction and modeling.

[0086] S3. Perform feature extraction on the processed meteorological data to obtain complexity feature information and dynamic feature information of the meteorological conditions corresponding to the various meteorological data.

[0087] Specifically, since the processed meteorological data integrates meteorological data from a variety of different sources, it can comprehensively and accurately capture the dynamic changes of meteorological conditions. The disclosed embodiment extracts important information that can reflect the complexity and dynamic characteristics of meteorological conditions by performing feature extraction on the processed meteorological data, so as to provide strong support for subsequent modeling and meteorological scene generation.

[0088] S4. Constructing a random evolution model of the meteorological conditions according to the complexity characteristic information and dynamic characteristic information of the meteorological conditions.

[0089] Specifically, according to the complexity and dynamic characteristics of meteorological conditions, the obtained random evolution module can describe the evolution process of meteorological conditions. Among them, meteorological data from different sources correspond to a variety of different meteorological conditions. The random evolution model can not only describe the average change trend of meteorological variables, but also capture random fluctuations in meteorological conditions. The random evolution model constructed by the embodiment of the present disclosure can simulate the evolution path of meteorological conditions at different time points, providing a basis for subsequent meteorological scene generation.

[0090] S5. Sampling the random evolution model multiple times to generate multiple different meteorological scenarios.

[0091] Specifically, the random evolution model can simulate the randomness of meteorological conditions. The disclosed embodiment samples the random evolution model, so that the embodiment can generate meteorological scenarios that reflect the diversity and uncertainty of meteorological conditions, laying a solid foundation for the generation of new energy output scenarios.

[0092] S6. Predicting the output of the renewable energy based on a plurality of different meteorological scenarios.

[0093] Specifically, the generation of renewable energy output scenarios depends on the dynamic changes of meteorological conditions, including key factors such as wind speed, solar radiation intensity and temperature. The goal of this step is to convert meteorological scenarios into specific renewable energy power output, providing direct input data for the dispatch and planning of the power system.

[0094] The disclosed embodiment can effectively integrate various types of meteorological information to form a high-quality, widely covered data set by fusing multi-source data from ground meteorological stations, satellite remote sensing and meteorological radar. The fusion of multi-source data overcomes the limitations of a single data source, making the description of meteorological conditions more comprehensive and accurate, and providing a solid data foundation for subsequent modeling and scene generation. Further, the disclosed embodiment solves the problem of resolution differences in time and space of multi-source meteorological data by performing time and space processing on various types of meteorological data, ensuring the consistency of data on a unified space-time grid. The processed data not only improves the accuracy of model input, but also avoids errors caused by data inconsistency, ensuring the reliability of subsequent feature extraction and modeling. Further, the disclosed embodiment extracts the complexity and nonlinear dynamic behavior of meteorological conditions, constructs a random evolution model, captures the random volatility of meteorological conditions, and then generates meteorological scenes that can cover a wide range of changes in meteorological conditions based on meteorological conditions, providing diversified inputs for new energy output prediction, enhancing the system's ability to cope with uncertainty, and making the final prediction results more accurate.

[0095] In some embodiments, step S2, performing temporal and spatial processing on the at least two types of meteorological data to obtain processed meteorological data, comprises the following steps:

[0096] S21. Unify the at least two meteorological data to the same time resolution, and fill in the missing data in the at least two meteorological data due to the asynchronization of data collection time, so as to obtain the time-aligned meteorological data.

[0097] Specifically, in the disclosed embodiment, since meteorological data from different sources usually have different time resolutions, such as the data from ground meteorological stations may have a time resolution of 10 minutes to 1 hour, the time resolution of satellite remote sensing data may be 15 minutes to 1 hour, and the time resolution of meteorological radar data may be several minutes to 15 minutes. Therefore, it is first necessary to unify the time resolution of all meteorological data to the same time resolution, for example, unify the time resolution of all meteorological data to the hourly level, which is a time scale commonly used in new energy output forecasting.

[0098] After unifying the time resolution of all meteorological data to the same time resolution, the missing time point data can be filled in based on the data corresponding to the adjacent time points.

[0099] In some embodiments, a linear interpolation method is used to fill in the missing data in the at least two meteorological data due to the asynchronous data collection time. Specifically, for the missing time point data, the data of the adjacent time points before and after are used for linear interpolation, and the formula is as follows:

[0100]

[0101] Wherein, t is the time point of missing data, t1 and t2 are adjacent time points of known data, X(t) is the filled data, and X(t1) and X(t2) are the values ​​of the meteorological data at t1 and t2, respectively. Through this method, it is ensured that all meteorological data are aligned at the same time point to form a temporally continuous meteorological data set. Of course, the embodiments of the present disclosure can also fill in the missing data in the at least two meteorological data due to the asynchronous data collection time according to other methods, and the present disclosure does not limit this.

[0102] S22, unifying the time-aligned meteorological data into the same spatial resolution, and filling in the missing data in the time-aligned meteorological data due to spatial asynchrony of data collection, to obtain the processed meteorological data.

[0103] Specifically, data from different data sources may also have differences in spatial resolution. For example, data from ground-based weather stations are usually unevenly distributed, satellite remote sensing data has a higher spatial resolution, and weather radar data is concentrated in a specific area. The disclosed embodiment can integrate these data into the same spatial grid through spatial interpolation, so that subsequent analysis and generation of new energy output scenarios are more accurate and comprehensive.

[0104] In some embodiments, filling missing data in the time-aligned meteorological data due to different spatial resolutions of the data in step S22 to obtain the processed meteorological data specifically includes:

[0105] S221. Calculate the spatial correlation between meteorological variables at different locations in the time-aligned meteorological data.

[0106] S222. Based on the spatial correlation, a semivariance function is constructed using the following formula.

[0107] where γ(h) represents the semivariance of the position pair with distance h, and Z(X i ) is the position X i where N(h) is the value of the meteorological variable at , and N(h) is the number of location pairs with a distance h.

[0108] S223, using the semivariogram function, calculating the weighted average of the interpolation points, and using the weighted average as the missing data in the time-aligned meteorological data due to the asynchrony of data collection space, to obtain the processed meteorological data. The value of is calculated by the following formula:

[0109] Among them, λ iis the weight, calculated based on the spatial position and semivariogram function, satisfying .

[0110] The disclosed embodiment ensures that the interpolation result can reflect the spatial relationship between the observation points through this method. Of course, the disclosed embodiment can also fill in the missing data in the time-aligned meteorological data due to the asynchrony of data collection space through other methods, and the disclosure does not limit this.

[0111] The disclosed embodiment integrates the processed meteorological data into a unified space-time grid by performing time synchronization and spatial interpolation processing on the at least two types of meteorological data. Through this integration, information from all data sources is integrated to form a high-quality meteorological data set with a uniform spatiotemporal resolution, providing a complete data foundation for subsequent feature extraction and model building. The disclosed embodiment successfully implements time synchronization and spatial interpolation processing of meteorological data, ensuring the consistency of multi-source data in time and space, thereby making subsequent analysis and generation of new energy output scenarios more accurate.

[0112] Figure 2 A flowchart of a specific execution process of step S3 provided in an embodiment of the present disclosure.

[0113] like Figure 2 As shown, in some embodiments, step S3, extracting features from the processed meteorological data to obtain complexity feature information of meteorological conditions corresponding to the at least two meteorological data, includes:

[0114] S31, dividing the time series of the processed meteorological data into different scales, and calculating the minimum number of boxes required to cover all the processed meteorological data at each scale.

[0115] S32. In a double logarithmic coordinate system, fit a complexity characteristic curve according to different division scales and the minimum number of boxes.

[0116] S33. Obtaining complexity characteristic information of meteorological conditions corresponding to the at least two meteorological data according to the complexity characteristic curve.

[0117] Specifically, suppose the scale of dividing the time series of processed meteorological data into , each scale The minimum number of boxes required to cover the entire dataset In the double logarithmic coordinate system, by fitting and The slope of the curve can be the complexity feature information D f , the formula is:

[0118]

[0119] The complexity feature information D f It can measure the complexity of data and reflect the self-similarity and detailed structure of meteorological data. f This was input into the subsequent modelling process to capture the complexity of meteorological data.

[0120] It is understandable that the embodiments of the present disclosure may also obtain the complexity characteristic information of the meteorological conditions corresponding to the various meteorological data by other methods, such as direct counting method, small vein method, etc., and the present disclosure does not limit this.

[0121] In some embodiments, step S3, extracting features from the processed meteorological data to obtain dynamic feature information of meteorological conditions corresponding to the multiple meteorological data, includes:

[0122] S34, mapping the one-dimensional time series in the time series of the processed meteorological data into a high-dimensional phase space.

[0123] Specifically, in this embodiment, in order to further capture the dynamic characteristics of meteorological data, a nonlinear dynamic modeling method is used. First, the meteorological time series is reconstructed in phase space. Specifically, the time delay embedding method can be used to map the one-dimensional time series X(t) into the high-dimensional phase space X(t), which is defined as:

[0124]

[0125] in, is the time delay, and m is the embedding dimension. The optimal embedding dimension m is selected by the False Nearest Neighbors (FNN) method, and the time delay is determined by the autocorrelation function .

[0126] S35. Calculate the exponential separation rate of adjacent points in the high-dimensional phase space, and determine the dynamic characteristic information of the meteorological conditions corresponding to the at least two meteorological data according to the exponential separation rate.

[0127] Specifically, the maximum Lyapunov exponent λ is calculated by calculating the exponential separation rate of adjacent points in the phase space trajectory. max :

[0128]

[0129] Where d(t) is the distance between adjacent points; if λ max,>0, the meteorological system exhibits chaotic characteristics, indicating the sensitive dependence of the system on the initial conditions. Among them, chaos reflects the sensitive dependence of the system and the uncertainty of long-term predictions, representing the dynamic characteristic information of meteorological conditions. The maximum Lyapunov exponent is input as a feature into the subsequent model to capture the chaotic characteristics and nonlinear dynamic behavior of meteorological data.

[0130] The disclosed embodiment extracts the complex features and dynamic features of meteorological data to obtain the self-similarity and detail structure of the meteorological data, as well as the uncertainty characteristics of the meteorological data, and inputs these features into subsequent meteorological condition modeling and new energy output scenario generation, so that the final scoring prediction results are more accurate and more in line with actual application scenarios.

[0131] Figure 3 A flowchart of executing step S4 is provided in an embodiment of the present disclosure.

[0132] like Figure 3 As shown, in some embodiments, step S4, constructing the random evolution model of the meteorological condition according to the complexity characteristic information and dynamic characteristic information of the meteorological condition, specifically includes:

[0133] S41. According to the complexity characteristic information and dynamic characteristic information of the meteorological conditions, statistical analysis is performed on the time series of the at least two meteorological data to determine the parameters of the drift term and the expansion term.

[0134] Specifically, the drift term represents the average change trend of meteorological variables in meteorological data, which is usually determined based on the statistical characteristics of the acquired historical meteorological data; the diffusion term represents the random fluctuation of meteorological variables in meteorological data, capturing the uncertainty in the meteorological system.

[0135] In some embodiments, based on the complexity characteristic information and dynamic characteristic information of the meteorological conditions, a statistical analysis is performed on the time series data of the at least two meteorological data to determine the parameters of the drift term and the extension term, specifically including: based on the complexity characteristic information and dynamic characteristic information of the meteorological conditions, a statistical analysis is performed on the time series data of the meteorological data to obtain the long-term average trend of the meteorological variables, and the parameters of the drift term and the extension term are determined based on the long-term average trend.

[0136] In some embodiments, according to the complexity characteristic information and dynamic characteristic information of the meteorological conditions, the time series of the at least two meteorological data are statistically analyzed to determine the parameters of the drift term and the expansion term, and specifically include: adjusting the parameters of the drift term and the expansion term by maximum likelihood estimation (MLE) so that the finally generated model has the best fitting effect on the at least two meteorological data obtained. The goal of the MLE method is to maximize the following likelihood function:

[0137]

[0138] Among them, θ is the model parameter, p(X(t i )|θ) is the probability density function of meteorological data under given parameters.

[0139] S42. Using meteorological variables as variables of the random evolution model, an initial random evolution model is obtained based on the parameters of the drift term and the extension term, wherein the drift term represents the average change trend of the meteorological variable, and the extension term represents the random fluctuation of the meteorological variable.

[0140] Specifically, the meteorological variable X(t) is modeled as a random evolution model to simultaneously consider the deterministic changes and random fluctuations of meteorological conditions. The basic form of this initial random evolution model is:

[0141] Dx(t)=μ(X(t),t)dt+σ(X(t),t)Dw(t)

[0142] Among them, μ(X(t), t) is the drift term, which represents the average change trend of meteorological variables and is usually determined based on the statistical characteristics of historical meteorological data; σ(X(t), t) is the diffusion term, which represents the random fluctuation of meteorological variables and captures the uncertainty in the meteorological system; W(t) is the standard Brownian motion, which represents the source of randomness in the model.

[0143] S43, discretizing the initial random evolution model to obtain the random evolution model to simulate the future evolution path of the meteorological conditions.

[0144] Specifically, the random evolution model is discretized to simulate the future evolution path of meteorological conditions, which specifically includes: discretizing the random evolution model using the following discretization formula:

[0145]

[0146] Among them, μ(X(t), t) is the drift term, σ(X(t), t) is the diffusion term, Δt is the time step, and Z is a standard normal distribution random variable. In actual operation, an appropriate time step Δt is selected and the discretization model is run multiple times to generate multiple possible meteorological evolution paths. These paths can reflect the possible states of meteorological conditions at different time points in the future and provide a basis for the diversity of meteorological scenarios.

[0147] Through the above steps S41 to S43, the embodiment of the present disclosure can construct an effective random evolution model, which can describe the random evolution process of meteorological conditions and generate multiple possible meteorological evolution paths through numerical simulation. These paths will serve as the basis for subsequent meteorological scene generation, helping to better understand and predict changes in new energy output.

[0148] After constructing and simulating the random evolution model of meteorological conditions, it is necessary to extract multiple representative meteorological scenarios from the random evolution model. These meteorological scenarios will be used for further prediction and optimization of renewable energy output. The Markov Chain Monte Carlo (MCMC) method is an effective sampling technique that can generate diverse samples from complex probability distributions. Through the MCMC method, multiple meteorological scenarios can be generated from the random evolution model to reflect different meteorological evolution paths.

[0149] In some embodiments, step S5, sampling the random evolution model at least twice to generate at least two different meteorological scenarios, includes:

[0150] S51. Select an initial meteorological state from the at least two meteorological data, and generate a candidate state according to the initial meteorological state.

[0151] Specifically, it is necessary to first set initial conditions for the sampling process. The initial meteorological state X0 can be selected from the acquired meteorological data to ensure that the initial state has realistic significance and conforms to the basic distribution of meteorological conditions. Then set the proposed distribution q (X'|Xt) to generate the candidate state X'. Optionally, the proposed distribution can be selected as a Gaussian distribution, because the Gaussian distribution has a simple mathematical form and is easy to calculate, which is convenient for random perturbations of the state of meteorological variables. Among them, the form of the proposed distribution is:

[0152]

[0153] in, Represents a Gaussian distribution with a mean of the current state Xt and a variance σ² that depends on the fluctuation range of the meteorological variable.

[0154] S52: Calculate the acceptance probability of the candidate state by comparing the posterior probability of the current state with the candidate state.

[0155] Specifically, in the embodiment, the acceptance probability α of the candidate state X' is calculated by comparing the posterior probabilities of the current state and the candidate state. The formula for the acceptance probability is:

[0156]

[0157] Among them, p(Xt|data) represents the posterior distribution of meteorological conditions supported by historical data, q(X'|Xt) is the proposed distribution, and Xt is the current state. The formula determines whether to accept a new state by comparing the relative "superiority" of the candidate state with the current state.

[0158] S53: Determine the current final state based on the acceptance probability and the randomly generated random number, and use the current final state as a sampling result.

[0159] Specifically, in each iteration, a random number U is first generated, which obeys the uniform distribution in the interval [0,1]. If U < α, the candidate state X' is accepted as the current meteorological condition state, otherwise the current state Xt is kept unchanged. This step ensures that the sampling process not only explores the high probability area of ​​meteorological conditions, but also retains the possibility of exploring the low probability area, increasing the diversity of sampling.

[0160] S54, repeating steps S51-S53, and generating at least two different meteorological scenes according to at least two sampling results.

[0161] Specifically, in this embodiment, a set of meteorological condition samples is generated through multiple iterations of sampling, and these samples represent various states of meteorological conditions that may occur in the future under the stochastic differential equation model.

[0162] In the embodiment of the present disclosure, multiple meteorological samples generated by the above sampling process are regarded as independent meteorological scenes. These meteorological scenes reflect different possible evolution paths of meteorological conditions, and each scene has a certain probability weight, representing the possible state of the meteorological system under specific conditions. Through these steps, the embodiment of the present disclosure successfully generates a variety of meteorological scenes from the stochastic differential equation model. These scenes cover a variety of possibilities of meteorological conditions and provide a rich data basis for the subsequent generation and optimization of new energy output scenes.

[0163] In some embodiments, S6, predicting the output of the new energy based on at least two different meteorological scenarios, includes: optimizing at least two objectives of at least two different meteorological scenarios to obtain an optimized meteorological scenario; and predicting the output of the new energy based on the optimized meteorological scenario.

[0164] Specifically, after generating multiple meteorological scenarios, these scenarios need to be optimized with multiple objectives to ensure the best balance between prediction accuracy, scenario diversity, and computational complexity. Multi-objective optimization can effectively improve the practicality of the scenarios, making these meteorological scenarios both highly accurate and able to cover possible changes in meteorological conditions while controlling the use of computing resources. By optimizing the generated meteorological scenarios, high-quality input data can be provided for subsequent new energy output forecasting and scheduling.

[0165] In some embodiments, the at least two different meteorological scenarios in step S6 are subjected to at least two objective optimizations to obtain an optimized meteorological scenario, including:

[0166] S611. Based on at least two different meteorological scenarios, determine at least two optimization objectives, and construct at least two objective optimization functions according to the at least two optimization objectives; wherein the optimization objectives include prediction accuracy, scenario diversity, and computational complexity.

[0167] Specifically, in this embodiment, multiple optimization objectives are defined for the generated meteorological scenario, wherein the multiple optimization objectives include:

[0168] Prediction accuracy: Evaluates how well the generated weather scenario fits the acquired weather data, quantified using the following error function:

[0169]

[0170] in, is the generated meteorological scene data, y i is the corresponding acquired meteorological data, and N is the number of generated meteorological scenes.

[0171] Scenario diversity: Evaluate the differences between multiple generated meteorological scenarios to ensure that they reflect the wide variation of meteorological conditions. Diversity is measured by the following formula:

[0172]

[0173] Among them, S j and S k are two generated meteorological scenarios, d(S j , S k ) is the difference measure between them, and M is the generated meteorological scene.

[0174] Computational complexity: Evaluate the computational cost required to generate the meteorological scenario to ensure that the optimization process does not exceed the predetermined computing resources. The computational complexity is quantified by the following formula:

[0175]

[0176] Among them, Time i is the computation time required to generate each meteorological scenario, and N is the number of meteorological scenarios generated.

[0177] S612: Optimize the at least two objective optimization functions to obtain an optimization result.

[0178] Specifically, in the embodiment of the present disclosure, a genetic algorithm can be used to optimize the defined multi-objective function. The genetic algorithm is a global search algorithm based on natural selection and genetic mechanism, and the specific steps are as follows:

[0179] Initialize the population: Randomly generate a set of initial model parameters, which are used to control the various steps of meteorological scene generation.

[0180] Fitness evaluation: Calculate the fitness of each individual, which is based on the comprehensive score of the above multiple objective functions. Fitness evaluation uses a linear weighted method to weight the scores of different objectives to form an overall fitness.

[0181] Selection operation: select parent individuals based on fitness values, using roulette or tournament selection. The selected individuals have a higher probability of entering the next generation.

[0182] Crossover and mutation operations: Perform crossover operations on the selected parent individuals to generate offspring individuals, and introduce random changes through mutation operations to increase the diversity of the population. The mutation probability and crossover probability are set according to the experimental conditions to balance the relationship between exploration and utilization.

[0183] S613. Determine a non-inferior solution in which the at least two optimization objectives in the optimization result are balanced, and select the optimal solution from the non-inferior solutions to ensure that the generated meteorological scenario achieves the best balance between prediction accuracy, scenario diversity and computational complexity, so as to obtain an optimized meteorological scenario.

[0184] Specifically, in multi-objective optimization, there are often situations where multiple objectives cannot be optimized to the optimal level at the same time. Therefore, in this embodiment, the Pareto frontier optimization method can be used to balance multiple objectives. Among them, the Pareto frontier optimization method is used to balance multiple objectives, specifically including:

[0185] Constructing the Pareto frontier: For each population, determine the non-inferior solutions that achieve a balance between different objectives and form the Pareto frontier. These non-inferior solutions represent the optimal solution for one objective without deteriorating the other objective.

[0186] Select Pareto optimal solution: Select the final optimal solution on the Pareto frontier to ensure that the generated meteorological scenario achieves the best balance between prediction accuracy, scenario diversity and computational complexity. These solutions are usually retained to the next generation for further optimization.

[0187] In this embodiment, after multiple iterations and optimizations, a set of multi-objective optimized meteorological scenarios are finally generated. These scenarios not only have good performance in prediction accuracy, but also cover a wide range of meteorological changes, and the computational complexity is effectively controlled.

[0188] Through the above steps, this embodiment successfully achieves multi-objective optimization of the generated meteorological scenarios, so that these scenarios have both high prediction accuracy and can reflect diverse meteorological conditions, while controlling the computing cost and ensuring the efficiency of the generation process.

[0189] After completing the multi-objective optimization of the meteorological scenario, it is necessary to generate new energy output scenarios based on these optimized meteorological scenarios. The generation of new energy output scenarios depends on the dynamic changes of meteorological conditions, including key factors such as wind speed, solar radiation intensity and temperature. The disclosed embodiment also needs to convert the meteorological scenario into a specific new energy power output to provide direct input data for the scheduling and planning of the power system.

[0190] In some embodiments, step S6, predicting the output of the new energy source based on at least two different meteorological scenarios, specifically includes:

[0191] S61. Extract key meteorological parameters from at least two different meteorological scenarios, input them into a preset new energy output prediction model, and calculate the power output of wind power generation and the power output of photovoltaic power generation, wherein the key meteorological data include wind speed data and solar radiation intensity data.

[0192] Specifically, the preset new energy output prediction model includes a preset wind power generation prediction model and a preset photovoltaic power generation prediction model.

[0193] The optimized weather scenarios are fed into pre-set forecasting models for wind and photovoltaic power generation, which calculate the power output of wind and photovoltaic power generation, respectively, based on the historical correlation between weather conditions and power generation.

[0194] For each optimized meteorological scenario, key meteorological parameters such as wind speed v(t), solar radiation intensity G(t) and ambient temperature T(t) are extracted as inputs to the new energy output prediction model.

[0195] In this embodiment, the wind speed power curve model is used to convert the wind speed data in the meteorological scene into the wind power output. The relationship between the wind power output Pw(t) and the wind speed v(t) is usually nonlinear, and the wind power output Pw(t) is calculated as follows:

[0196]

[0197] Among them, Vcut-in is the cut-in wind speed of the wind turbine, Vcut-out is the cut-out wind speed, Vrated is the rated wind speed, and Prated is the rated power of the wind turbine. Through this model, the corresponding wind power generation is calculated based on the change of wind speed.

[0198] In this embodiment, a power model based on solar radiation intensity is used to convert solar radiation data in meteorological scenes into photovoltaic power output. Photovoltaic power is related to factors such as solar radiation intensity and temperature, and its calculation formula is as follows:

[0199] Ppv(t)=η*A* G(t)*(1-β*(T(t))-Tref)

[0200] Among them, η is the efficiency of the photovoltaic module, A is the total area of ​​the photovoltaic array, G(t) is the solar radiation intensity, T(t) is the ambient temperature, Tref is the reference temperature, and β is the temperature coefficient, which reflects the impact of temperature on the efficiency of the photovoltaic module. The model can calculate the photovoltaic power generation under different meteorological conditions.

[0201] S62: Based on the power output of the wind power generation and the power output of the photovoltaic power generation, obtain total new energy output scenarios at different time points.

[0202] Specifically, in this embodiment, the wind power generation power Pw(t) and the photovoltaic power generation power Pwv(t) are combined to generate the total new energy output scenario Ptotal(t) at different time points. The calculation formula for the comprehensive new energy output is as follows:

[0203] Ptotal(t)=Pw(t)+Pwv(t)

[0204] This formula combines the power generation of wind and photovoltaic power to obtain the total new energy output scenario, reflecting the overall power generation capacity of new energy under given meteorological conditions.

[0205] S63: predicting the new energy output based on the total new energy output scenario.

[0206] In this embodiment, the generated renewable energy output scenarios can be directly applied to the dispatching and planning of the power system. These scenarios provide the power dispatching center with detailed predictions of future renewable energy generation, help optimize the allocation of power resources, and formulate reasonable dispatching strategies.

[0207] Through the above steps S61-S63, this embodiment converts the optimized meteorological scenario into a specific renewable energy output scenario. These output scenarios accurately reflect the renewable energy power generation under different meteorological conditions, providing key data support for the practical application of the power system.

[0208] The method for predicting the output of renewable energy proposed in the embodiment of the present disclosure, through time synchronization and spatial interpolation processing, advanced feature extraction, random evolution modeling, Markov chain Monte Carlo sampling, multi-objective optimization and other steps, generates renewable energy output scenarios with high precision, high diversity and controllable computational complexity, which can effectively cope with the complexity and uncertainty of meteorological conditions and provide important support for the stable operation of the power system. The method for predicting the output of renewable energy provided in the embodiment of the present disclosure has the following technical effects:

[0209] On the one hand, the present disclosure can effectively integrate various types of meteorological information to form a high-quality, wide-coverage data set by fusing multi-source data from ground meteorological stations, satellite remote sensing, and meteorological radar. The fusion of multi-source data overcomes the limitations of a single data source, making the description of meteorological conditions more comprehensive and accurate, and providing a solid data foundation for subsequent modeling and scenario generation.

[0210] On the other hand, the present disclosure solves the problem of resolution differences in time and space of multi-source meteorological data through time synchronization processing and spatial interpolation processing, ensuring the consistency of data on a unified time and space grid. The processed data not only improves the accuracy of model input, but also avoids errors caused by data inconsistency, ensuring the reliability of subsequent feature extraction and modeling.

[0211] On the other hand, the present disclosure can deeply capture the complexity and nonlinear dynamic behavior of meteorological conditions by extracting fractal characteristics and dynamic characteristics of meteorological data. These characteristics provide rich information for the construction of stochastic differential equation models, improving the model's adaptability to meteorological conditions and prediction accuracy.

[0212] On the other hand, the present disclosure successfully simulates the random evolution of meteorological conditions by constructing a stochastic differential equation model. The model not only takes into account the average change trend of meteorological variables, but also effectively captures the random volatility in the meteorological system, laying a theoretical foundation for the generation of meteorological scenes.

[0213] On the other hand, the present disclosure uses the Markov chain Monte Carlo method to sample from the stochastic differential equation model to generate a variety of meteorological scenarios. These scenarios can cover a wide range of changes in meteorological conditions, provide a variety of inputs for subsequent new energy output forecasts, and enhance the system's ability to cope with uncertainty.

[0214] On the other hand, the present disclosure ensures the best balance between prediction accuracy, scenario diversity and computational complexity by performing multi-objective optimization on the generated meteorological scenarios. The optimized meteorological scenarios are not only accurate but also well representative, while the computational cost is effectively controlled, ensuring the efficiency and practicality of the model.

[0215] On the other hand, the new energy output scenarios generated by the present disclosure based on the optimized meteorological scenarios can accurately reflect the power generation situation under different meteorological conditions. These output scenarios provide a reliable basis for the dispatch and planning of the power system, help optimize the configuration of power resources, and improve the operating efficiency and stability of the power system.

[0216] On the second aspect, based on the same inventive concept, the disclosed embodiment also provides a new energy output prediction device.

[0217] Figure 4 FIG. 1 is a structural block diagram of a new energy output prediction device 100 provided in an embodiment of the present disclosure. Figure 4 As shown, the prediction device 100 includes a data collection module 101, a data processing module 102, a feature extraction module 103, a random modeling module 104, a scene generation module 105 and a prediction module 106.

[0218] The data acquisition module 101 is used to obtain a variety of meteorological data from different sources, and the various meteorological data at least include ground meteorological station data, satellite remote sensing data and meteorological radar data.

[0219] The data processing module 102 is used to perform temporal and spatial processing on the various meteorological data to obtain processed meteorological data.

[0220] The feature extraction module 103 is used to extract features from the processed meteorological data to obtain complexity feature information and dynamic feature information of meteorological conditions corresponding to the various meteorological data.

[0221] The stochastic modeling module 104 is used to construct a stochastic evolution model of the meteorological condition according to the complexity characteristic information and dynamic characteristic information of the meteorological condition.

[0222] The scenario generation module 105 is used to perform multiple sampling on the random evolution model to generate multiple different meteorological scenarios.

[0223] The prediction module 106 is used to predict the output of the new energy based on a plurality of different meteorological scenarios.

[0224] In some embodiments, the ground meteorological station data includes at least one of the following: ground temperature data, humidity data, wind speed data, and air pressure data; the satellite remote sensing data includes at least one of the following: solar radiation data, cloud cover data, and surface temperature; the meteorological radar data includes precipitation intensity data and / or wind speed distribution data.

[0225] In some embodiments, the data processing module 102 includes a time processing module 1021 and a space processing module 1022. The time processing module 1021 is used to unify the at least two meteorological data to the same time resolution, and fill in the missing data in the at least two meteorological data due to the asynchronous data collection time, so as to obtain the time-aligned meteorological data. The space processing module 1022 is used to unify the time-aligned meteorological data to the same spatial resolution, and fill in the missing data in the time-aligned meteorological data due to the different data spatial resolutions, so as to obtain the processed meteorological data.

[0226] Figure 5 A structural block diagram of a feature extraction module 103 provided in an embodiment of the present disclosure.

[0227] like Figure 5 As shown, in some embodiments, the feature extraction module 103 includes a complexity feature extraction module 31 and a dynamic feature extraction module 32 , wherein the complexity feature extraction module 31 includes a division module 311 , a fitting module 312 and an extraction module 313 .

[0228] The division module 311 is used to divide the time series of the processed meteorological data into different scales and calculate the minimum number of boxes required to cover all the processed meteorological data at each scale. The fitting module 312 is used to fit the complexity characteristic curve according to the different scales of division and the minimum number of boxes in a double logarithmic coordinate system. The extraction module 313 is used to obtain the complexity characteristic information of the meteorological conditions corresponding to the at least two meteorological data according to the complexity characteristic curve.

[0229] In some embodiments, the dynamic feature extraction module 32 includes a mapping module 321 and a first determination module 322. The mapping module 321 is used to map the one-dimensional time series in the time series of the processed meteorological data to the high-dimensional phase space. The first determination module 322 is used to calculate the exponential separation rate of adjacent points in the high-dimensional phase space, and determine the dynamic feature information of the meteorological conditions corresponding to the at least two meteorological data according to the exponential separation rate.

[0230] In some embodiments, the stochastic modeling module 104 includes a statistical module 1041, a random module 1042 and a discrete module 1043. The statistical module 1041 is used to perform statistical analysis on the time series of the at least two meteorological data according to the complexity characteristic information and dynamic characteristic information of the meteorological conditions, and determine the parameters of the drift term and the extension term. The random module 1042 is used to use meteorological variables as variables of the stochastic evolution model, and obtain an initial stochastic evolution model based on the parameters of the drift term and the extension term, wherein the drift term represents the average change trend of the meteorological variable, and the extension term represents the random fluctuation of the meteorological variable. The discrete module 1043 is used to discretize the initial stochastic evolution model to obtain the stochastic evolution model to simulate the future evolution path of the meteorological conditions.

[0231] In some embodiments, the scene generation module 105 includes a selection module 1051, a comparison module 1052, a sampling module 1053 and a meteorological scene module 1054. Among them, the selection module 1051 is used to select an initial meteorological state from the at least two meteorological data, and generate a candidate state based on the initial meteorological state. The comparison module 1052 is used to calculate the acceptance probability of the candidate state by comparing the posterior probability of the current state with the candidate state. The sampling module 1053 is used to determine the current final state based on the acceptance probability and the randomly generated random number, and use the current final state as a sampling result. The meteorological scene module 1054 is used to generate at least two different meteorological scenes based on at least two sampling results.

[0232] Figure 6 A structural block diagram of a prediction module 106 provided in an embodiment of the present disclosure.

[0233] like Figure 6 As shown, in some embodiments, the prediction module 106 includes a calculation module 1061, an output generation module 1062 and a prediction submodule 1063. The calculation module 1061 is used to extract key meteorological parameters in at least two different meteorological scenarios, input them into a preset new energy output prediction model, and calculate the power output of wind power generation and the power output of photovoltaic power generation, and the key meteorological data include wind speed data and solar radiation intensity data. The output generation module 1062 is used to generate total new energy output scenarios at different time points based on the power output of wind power generation and the power output of photovoltaic power generation. The prediction submodule 1063 is used to predict the new energy output based on the total new energy output scenario.

[0234] In some embodiments, the prediction module 106 further includes an optimization module 1064, wherein the optimization module 1064 is used to perform multi-objective optimization on a plurality of different meteorological scenarios to obtain an optimized meteorological scenario. The calculation module 1061 is specifically used to extract key meteorological parameters in the optimized meteorological scenario, input them into a preset new energy output prediction model, and calculate the power output of wind power generation and the power output of photovoltaic power generation.

[0235] In some embodiments, the optimization module 1064 includes a construction module 641, an optimization submodule 642, and a balancing module 643. The construction module 641 is used to determine at least two optimization objectives based on at least two different meteorological scenarios, and to construct at least two target optimization functions according to the at least two optimization objectives; wherein the optimization objectives include prediction accuracy, scenario diversity, and computational complexity. The optimization submodule 642 is used to optimize the at least two target optimization functions to obtain an optimization result. The balancing module 643 is used to determine a non-inferior solution in which the at least two optimization objectives in the optimization result are balanced, and to select the optimal solution from the non-inferior solution to ensure that the generated meteorological scenario achieves the best balance between prediction accuracy, scenario diversity, and computational complexity, and obtain an optimized meteorological scenario.

[0236] It should be noted that the specific details of each module in the prediction device 100 provided in the embodiment of the present disclosure can refer to the specific embodiment of the above-mentioned prediction method, and the present disclosure will not repeat them.

[0237] Figure 7 FIG. 7 is a structural block diagram of an electronic device 70 provided in an embodiment of the present disclosure. Figure 7 As shown, the electronic device 70 includes a memory 72, a processor 71, and a computer program stored in the memory 72 and executable on the processor 71. When the processor 71 executes the computer program, any method described in the first aspect is implemented.

[0238] In some embodiments, the electronic device 70 may further include a computer storage device 73 for storing a computer program of any of the methods described in the first aspect above.

[0239] In a fourth aspect, an embodiment of the present disclosure further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program of the method described in any one of the first aspects above.

[0240] Among them, the computer-readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable red-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, disk or optical disk.

[0241] In a fifth aspect, an embodiment of the present disclosure further provides a computer program product, comprising a computer program / instructions, which, when executed by a processor, implement the steps of any one of the methods described in the first aspect above.

[0242] Those skilled in the art will appreciate that the embodiments of the present disclosure may be provided as methods, systems, or computer program products. Therefore, the present disclosure may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present disclosure may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0243] The present disclosure is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present disclosure. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0244] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0245] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0246] Specific embodiments are used in the present disclosure to illustrate the principles and implementation methods of the present disclosure. The description of the above embodiments is only used to help understand the method of the present disclosure and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present disclosure, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present disclosure.

Claims

1. A method for predicting the output of new energy, characterized in that: The method comprises: Acquire at least two types of meteorological data from different sources, wherein the meteorological data includes at least two of ground meteorological station data, satellite remote sensing data, and meteorological radar data; Performing temporal and spatial processing on the at least two types of meteorological data to obtain processed meteorological data; Performing feature extraction on the processed meteorological data to obtain complexity feature information and dynamic feature information of meteorological conditions corresponding to the at least two types of meteorological data; Constructing a random evolution model of the meteorological conditions according to the complexity characteristic information and dynamic characteristic information of the meteorological conditions; Sampling the random evolution model at least twice to generate at least two different meteorological scenarios; Based on at least two different meteorological scenarios, the output of the new energy is predicted; wherein, The step of extracting features from the processed meteorological data to obtain complexity feature information of meteorological conditions corresponding to the at least two types of meteorological data includes: Dividing the time series of the processed meteorological data into different scales, and calculating the minimum number of boxes required to cover all the processed meteorological data at each scale; In a double logarithmic coordinate system, a complexity characteristic curve is fitted according to different scales of division and the minimum number of boxes; Obtaining complexity characteristic information of meteorological conditions corresponding to the at least two meteorological data according to the complexity characteristic curve; The step of extracting features from the processed meteorological data to obtain dynamic feature information of meteorological conditions corresponding to the at least two types of meteorological data includes: Mapping the one-dimensional time series in the time series of the processed meteorological data into a high-dimensional phase space; The exponential separation rate of adjacent points in the high-dimensional phase space is calculated, and the dynamic characteristic information of the meteorological conditions corresponding to the at least two meteorological data is determined according to the exponential separation rate.

2. The method according to claim 1, characterized in that The ground weather station data includes at least one of the following: ground temperature data, humidity data, wind speed data, and air pressure data; The satellite remote sensing data includes at least one of the following: solar radiation data, cloud cover data, and surface temperature data; The weather radar data includes precipitation intensity data and / or wind speed distribution data.

3. The method according to claim 1, characterized in that Processing the at least two types of meteorological data in time and space to obtain processed meteorological data includes: Unifying the at least two meteorological data to the same time resolution, and filling in missing data in the at least two meteorological data due to asynchronous data collection time, to obtain time-aligned meteorological data; The time-aligned meteorological data are unified into the same spatial resolution, and missing data in the time-aligned meteorological data due to different spatial resolutions of data are filled to obtain the processed meteorological data.

4. The method according to claim 1, characterized in that: The step of constructing the random evolution model of the meteorological conditions according to the complexity characteristic information and the dynamic characteristic information of the meteorological conditions comprises: According to the complexity characteristic information and dynamic characteristic information of the meteorological conditions, statistical analysis is performed on the time series of the at least two meteorological data to determine the parameters of the drift term and the expansion term; Taking meteorological variables as variables of the random evolution model, and based on the parameters of the drift term and the extension term, an initial random evolution model is obtained, wherein the drift term represents the average change trend of the meteorological variable, and the extension term represents the random fluctuation of the meteorological variable; The initial random evolution model is discretized to obtain the random evolution model to simulate the future evolution path of the meteorological conditions.

5. The method according to claim 1, characterized in that The random evolution model is sampled at least twice to generate at least two different meteorological scenarios, including: Selecting an initial meteorological state from the at least two meteorological data, and generating a candidate state according to the initial meteorological state; By comparing the posterior probability of the current state with the candidate state, calculating the acceptance probability of the candidate state; Determine the current final state based on the acceptance probability and the randomly generated random number, and use the current final state as a sampling result; At least two different meteorological scenarios are generated based on at least two sampling results.

6. The method according to claim 1, characterized in that The forecasting of the output of the renewable energy source based on at least two different meteorological scenarios includes: Performing at least two objective optimizations on at least two different meteorological scenarios to obtain an optimized meteorological scenario; The output of the renewable energy is predicted based on the optimized meteorological scenario.

7. The method according to claim 6, characterized in that The at least two different meteorological scenarios are subjected to at least two objective optimizations to obtain an optimized meteorological scenario, including: Based on at least two different meteorological scenarios, at least two optimization objectives are determined, and at least two objective optimization functions are constructed according to the at least two optimization objectives; wherein the optimization objectives include prediction accuracy, scenario diversity, and computational complexity; Optimizing the at least two objective optimization functions to obtain an optimization result; Determine a non-inferior solution in which the at least two optimization objectives in the optimization result are balanced, and select the optimal solution from the non-inferior solutions to ensure that the generated meteorological scenario achieves the best balance between prediction accuracy, scenario diversity and computational complexity, so as to obtain an optimized meteorological scenario.

8. The method according to claim 1, characterized in that The forecasting of the output of the renewable energy source based on at least two different meteorological scenarios includes: Extract key meteorological parameters from at least two different meteorological scenarios, input them into a preset new energy output prediction model, and calculate the power output of wind power generation and the power output of photovoltaic power generation, wherein the key meteorological parameters include wind speed data and solar radiation intensity data; Based on the power output of the wind power generation and the power output of the photovoltaic power generation, a total new energy output scenario at different time points is obtained; The new energy output is predicted based on the total new energy output scenario.

9. A prediction device for new energy output, characterized in that: The device comprises: A data acquisition module, used to obtain a variety of meteorological data from different sources, wherein the multiple meteorological data at least include ground meteorological station data, satellite remote sensing data and meteorological radar data; A data processing module, used for performing temporal and spatial processing on the multiple meteorological data to obtain processed meteorological data; A feature extraction module is used to extract features from the processed meteorological data to obtain complex feature information and dynamic feature information of meteorological conditions corresponding to the various meteorological data; A stochastic modeling module, used for constructing a stochastic evolution model of the meteorological conditions according to the complexity characteristic information and dynamic characteristic information of the meteorological conditions; A scenario generation module, used to perform multiple sampling on the random evolution model to generate multiple different meteorological scenarios; A prediction module, used for predicting the output of the renewable energy based on a plurality of different meteorological scenarios; The feature extraction module includes a complexity feature extraction module and a dynamic feature extraction module, wherein the complexity feature extraction module includes a division module, a fitting module and an extraction module; The partitioning module is used to partition the time series of the processed meteorological data into different scales, and calculate the minimum number of boxes required to cover all the processed meteorological data at each scale; The fitting module is used for fitting a complexity characteristic curve in a double logarithmic coordinate system according to different scales of division and the minimum number of boxes; The extraction module is used to obtain complexity characteristic information of meteorological conditions corresponding to at least two meteorological data according to the complexity characteristic curve; The dynamic feature extraction module includes a mapping module and a first determination module, wherein: The mapping module is used to map the one-dimensional time series in the time series of the processed meteorological data into a high-dimensional phase space; The first determination module is used to calculate the exponential separation rate of adjacent points in the high-dimensional phase space, and determine the dynamic characteristic information of the meteorological conditions corresponding to the at least two meteorological data according to the exponential separation rate.

10. The device according to claim 9, characterized in that The data processing module comprises: A time processing module, used to unify the at least two meteorological data to the same time resolution, and fill in the missing data in the at least two meteorological data due to the asynchronization of data collection time, so as to obtain the time-aligned meteorological data; The spatial processing module is used to unify the time-aligned meteorological data into the same spatial resolution and fill in the missing data in the time-aligned meteorological data due to different spatial resolutions of the data to obtain the processed meteorological data.

11. The device according to claim 9, characterized in that The prediction module comprises: A calculation module is used to extract key meteorological parameters in at least two different meteorological scenes, input them into a preset new energy output prediction model, and calculate the power output of wind power generation and the power output of photovoltaic power generation, wherein the key meteorological parameters include wind speed data and solar radiation intensity data; An output generation module, used to generate total renewable energy output scenarios at different time points based on the power output of the wind power generation and the power output of the photovoltaic power generation; A prediction submodule is used to predict the new energy output based on the total new energy output scenario.

12. The device according to claim 11, characterized in that The prediction module also includes: An optimization module is used to perform multi-objective optimization on multiple different meteorological scenarios to obtain optimized meteorological scenarios; The calculation module is specifically used to extract key meteorological parameters in the optimized meteorological scene, input them into a preset new energy output prediction model, and calculate the power output of wind power generation and the power output of photovoltaic power generation.

13. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 8 is implemented.

14. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program for executing the method according to any one of claims 1 to 8.

15. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images