Hierarchical double-model fused wind power prediction data model

Abstract

The invention discloses a hierarchical double-model fused wind power prediction data model, and relates to a digital intelligent technology in a wind power scenario, and the data model firstly constructs a data layer based on a Transform architecture and a convolutional neural network, then carries out the pre-training of the data layer through a pre-constructed annotation data set, and finally carries out the pre-training of the data layer. Obtaining a double-model fused wind power prediction data model, and recording the model as a model A; and an independent data layer is further constructed on the basis of the model A and is recorded as a model B, so that a hierarchical data model architecture is formed. The hierarchical model architecture provided by the invention is pioneering in the technical field, and the efficiency of data intelligent modeling of the wind power can be remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of digital intelligent technology in wind power scenarios, and in particular to a hierarchical dual-model fusion wind power prediction data model. Background Art

[0002] Renewable energy power generation, especially wind power and photovoltaic power generation, as the main force of renewable energy power generation, has developed rapidly in recent years. However, since renewable energy power generation is affected by many meteorological factors, such as wind speed, light intensity, temperature, etc., the power generation has strong volatility and uncertainty. This intermittent and uncontrollable nature has brought great challenges to the dispatching and stable operation of the power grid. With the continuous increase in the capacity of renewable energy grid connection, how to effectively predict the power generation of renewable energy has become a key link in power grid planning and dispatching. Accurate power prediction can not only provide guarantee for the safe and stable operation of the power grid, but also optimize the grid dispatching, reduce spare capacity and dispatching costs, and thus improve the economic benefits of renewable energy power generation. At present, the new energy power prediction methods are mainly divided into traditional statistical methods and intelligent prediction methods based on artificial intelligence. Traditional methods such as time series analysis and regression models rely on the trend extrapolation of historical data, but perform poorly when dealing with complex nonlinear relationships. Intelligent prediction methods such as neural networks and support vector machines, although they can model nonlinear relationships to a certain extent, are limited by the complexity of the algorithm, the amount of data required, and the ability to capture long-term dependencies, and their prediction effects are not ideal.

[0003] In recent years, with the rapid development of deep learning technology, deep learning models such as convolutional neural networks (CNN) and Transformer have demonstrated powerful capabilities in time series data processing and feature extraction. CNN is good at extracting short-term dependency features from local data and can effectively capture local fluctuation patterns in renewable energy power data; while Transformer can model dependencies over a long period of time through the self-attention mechanism, which is suitable for capturing global features in power forecasting. However, prediction methods that use only CNN or Transformer often find it difficult to simultaneously extract local and global features. Therefore, our technical team considered fusing the two to obtain a wind power data model with inherent multi-features.

[0004] In particular, further examining the power algorithms of various existing wind power and other new energy sources, no matter how complex their internal structure or the specific form of their data algorithms are, they are essentially single-layer data architectures. The introduction and construction of multi-layer data architectures - especially considering the temporal characteristics of meteorological and geographical factors in wind power technology, and the time sequence characteristics of different geographical locations experiencing similar meteorological conditions, inspired our technical team to build a new two-layer (or even multi-layer) data architecture, which will also bring beneficial performance improvements to the digital intelligent modeling of wind power. Summary of the invention

[0005] The technical problem to be solved by the present invention is to overcome the various shortcomings of the prior art and provide a hierarchical dual-model fusion wind power prediction data model. This hierarchical data model not only considers the fusion of CNN or Transformer, but also especially considers the temporal characteristics of meteorological and geographical factors in wind power technology, and the time sequence of wind power scenes in different geographical locations experiencing similar meteorological conditions.

[0006] A hierarchical dual-model fusion wind power prediction data model builds a data layer based on the Transformer architecture and convolutional neural network, and pre-trains the data layer using a pre-constructed labeled data set to obtain a dual-model fusion wind power prediction data model, denoted as model A; further, an independent data layer is constructed on the basis of model A, denoted as model B, thereby forming a hierarchical data model architecture; wherein the model A serves as the lower-layer data model, the model B serves as the upper-layer data model, and the model B optimizes all adjustable parameters in the model A.

[0007] As a preferred technical solution of the present invention, the B model is constructed as a data model integrating the following two data elements: element x: meteorological spatiotemporal dynamic characteristics; element y: wind power generation scenario.

[0008] As a preferred technical solution of the present invention, the correlation relationship between element x and element y is: element x considers the spatiotemporal order of the impact of the spatiotemporal dynamic characteristics of meteorological conditions on different elements y; element x affects wind power generation scenarios in different geographical locations in sequence according to the spatiotemporal order, and the different wind power generation scenarios affected successively correspond to an ordered data set.

[0009] As a preferred technical solution of the present invention, the B model is constructed as the upper structure of the A model. First, for the pre-trained A model, the B model pre-adjusts the adjustable parameters therein, and obtains the variant model set A′ after the pre-adjustment; then, each element in the variant model set A′ acts on each element in the ordered data set corresponding to the element y; since the element x acts on each element of the element y in sequence according to the time and space order, the B model can perform data observation, that is, after observing that each element in the variant model set A′ acts on the first few elements of y, one or several variant models whose prediction effect is closest to the actual wind power data are identified, and a new variant model set is obtained, denoted as A′′, which includes the optimized A model with better prediction efficiency in the current cold wave weather; in this way, the A model is adjusted in real time based on A′′, so as to realize the real-time dynamic optimization of wind power prediction for the last few elements in the ordered data set corresponding to the element y.

[0010] As a preferred technical solution of the present invention, the A model integrates the multi-level feature extraction of Transformer and convolutional neural network to construct a process-based data structure to obtain a wind power prediction data model integrated with the two models; the process-based data structure of the A model includes: data process structure one, preprocessing the historical wind power data and related meteorological data and generating an input feature matrix; data process structure two, using the convolutional neural network CNN to extract local features of the preprocessed input data, capturing the short-term dependencies in the wind power time series and generating a preliminary feature representation; data process structure three, inputting the preliminary feature representation in the data process structure two into the Transformer model, extracting long-term time-dependent features through the self-attention mechanism and enhancing the global representation ability of the features; data process structure four, hierarchically integrating the local features extracted by CNN and the global features extracted by Transformer to generate a comprehensive feature representation; data process structure five, performing multi-step prediction of wind power based on the fused features and using a fully connected layer to output the final prediction result.

[0011] As a preferred technical solution of the present invention, the convolutional neural network in the data process structure 2 adopts a multi-layer convolution structure to abstract the local features in the time series layer by layer and use the pooling layer to perform feature dimensionality reduction.

[0012] As a preferred technical solution of the present invention, the Transformer model in the data process structure three adopts a multi-layer encoder structure. Each layer of the encoder includes a self-attention mechanism and a feedforward neural network. The multi-head self-attention mechanism is used to capture the dependency between different time steps in the input sequence, and the global information of the sequence is captured by calculating the correlation weights between each time step and other time steps. At the same time, the position information in the time series is explicitly encoded into the input features in combination with position encoding.

[0013] As a preferred technical solution of the present invention, it is characterized in that: the multi-level feature fusion in the data process structure four effectively fuses the local features extracted by CNN and the global features extracted by Transformer by weighted summation or splicing; CNN extracts short-term dependency features from the local window and captures the local patterns and detail features in the wind power time series, and Transformer can extract global information and model dependencies over a long period of time through a self-attention mechanism; the accuracy of the prediction is improved by fusing the features of the two and utilizing short-term and long-term information at the same time; in the process of feature fusion, the weights are dynamically adjusted according to the importance of local and global features by weighted summation, or feature splicing is used to retain more original information; in addition, an additional attention mechanism is introduced after fusion to improve the model's sensitivity to wind power fluctuations and prediction accuracy by learning the correlation between different features (focusing on features that are more important to the prediction task).

[0014] As a preferred technical solution of the present invention, the prediction part in the data process structure five adopts a multi-step prediction algorithm structure, which predicts the wind power changes in multiple time steps in the future by setting multiple output nodes at the same time, avoiding the problem of error accumulation in step-by-step prediction and directly giving the power prediction results of multiple time points in the future within one prediction cycle; each output node corresponds to the predicted value of a specific time step in the future, and the model learns the dependency between each time step in the time series through training and then makes a comprehensive prediction of the power trend in the future time period.

[0015] After adopting the above technical scheme, the beneficial technical effects achieved by the present invention are as follows: In essence, the technical team of the present invention has carried out technical innovations on two levels. First, a multi-level feature extraction algorithm integrating CNN and Transformer is proposed. Combining the advantages of both CNN and Transformer can more comprehensively model the timing characteristics of new energy power and improve prediction accuracy and reliability; second, the introduction and construction of a two-layer data architecture, which is the core of the technical development of the present invention. The timing characteristics of wind power generation under meteorological and geographical conditions inspired our technical team to build a new two-layer data architecture. This hierarchical model architecture is the first in the industry and is expected to bring significant efficiency improvements to the digital intelligent modeling of wind power. Further technical effects of the present invention are also described in detail in Examples 1 and 3 below. DETAILED DESCRIPTION

[0016] The specific implementation of the present invention is further described in detail below in conjunction with the implementation examples. The following implementation examples are used to illustrate the present invention, but are not used to limit the scope of the present invention.

[0017] In the description of the following embodiments, specific details such as specific system structures, technologies, etc. are provided for illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

[0018] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or combinations thereof.

[0019] It should also be understood that the term “and / or” used in the specification and appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0020] As used in the specification and appended claims of this application, the term "if" can be interpreted as "when" or "uponce" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "uponce it is determined" or "in response to determining" or "uponce [described condition or event] is detected" or "in response to detecting [described condition or event]", depending on the context.

[0021] In addition, in the description of the present application specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0022] References to "one embodiment" or "some embodiments" etc. described in the specification of this application mean that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the statements "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0023] Example 1: Wind power prediction data model based on dual-model fusion (Model A) Data model A is the bottom layer (lower layer) data model, which is constructed as a dual-model data structure that integrates Transformer and convolutional neural network, which is what we call the wind power prediction data model of dual-model fusion. In summary, the data structure of model A includes: a. Preprocessing historical wind power data and related meteorological data, including data cleaning, normalization and missing value processing; b. Using convolutional neural network (CNN) to extract local features of preprocessed input data, capture short-term dependencies in wind power time series, and generate preliminary feature representation; c. Input the preliminary feature representation in b into the Transformer model, extract long-term time dependency features through the self-attention mechanism, and enhance the global representation ability of features; d. Multi-level feature fusion of local features extracted by CNN and global features extracted by Transformer to generate comprehensive feature representation; e. Multi-step prediction of wind power based on fused features, and output the final prediction result using the fully connected layer.

[0024] Among them, the Transformer model in c adopts a multi-layer encoder structure. Each layer of the encoder includes a self-attention mechanism and a feedforward neural network. The multi-head self-attention mechanism is used to capture the dependencies between different time steps in the input sequence, which can effectively model the long-range dependency information in the wind power time series; the self-attention mechanism calculates the correlation weight between each time step and other time steps, thereby capturing the global information of the sequence, which can overcome the problem that traditional recurrent neural networks are difficult to handle long-distance dependencies; at the same time, combined with position encoding, the position information in the time series is explicitly encoded into the input features, which can make up for the lack of sequential information in the Transformer model itself, so that the model can identify the changing trend in the time sequence and enhance the accuracy and stability of wind power time series prediction.

[0025] Among them, the multi-level feature fusion in d effectively fuses the local features extracted by CNN and the global features extracted by Transformer through weighted summation or splicing; CNN can extract short-term dependency features from local windows and capture local patterns and detail features in wind power time series, and Transformer can extract global information and model dependencies over a long period of time through the self-attention mechanism; by fusing the features of the two, both short-term and long-term information can be used to improve the accuracy of prediction; in the process of feature fusion, a weighted summation method can be used to dynamically adjust the weights according to the importance of local and global features, or feature splicing can be used to retain more original information; in addition, an additional attention mechanism is introduced after fusion to further optimize the fusion effect, and by learning the correlation between different features, the focus is on the features that are more important to the prediction task, thereby improving the model's sensitivity to wind power fluctuations and the accuracy of prediction.

[0026] Among them, the prediction part in e adopts a multi-step prediction strategy. By setting multiple output nodes, the model can simultaneously predict the changes in wind power in multiple time steps in the future; this multi-step prediction strategy can not only avoid the problem of error accumulation in step-by-step prediction, but also directly give the power prediction results of multiple time points in the future within one prediction cycle, which greatly improves the timeliness and practicality of the model; each output node corresponds to the prediction value of a specific time step in the future. Through training, the model can learn the complex dependencies between the time steps in the time series, and then make a comprehensive and accurate prediction of the power trend in the future period; this strategy not only improves the prediction efficiency, but also makes the model more adaptable, able to cope with the fluctuations of wind power on different time scales, and enhance the feedback and guidance on the actual operation of the wind farm.

[0027] Example 2: Data execution process of model A in application scenario Model A is a data model based on Transformer and convolutional neural network (CNN) for wind power and other renewable energy power forecasting. From the perspective of data execution, it mainly includes the following steps: S1: Collect historical power data and related meteorological data of wind power renewable energy power stations. Meteorological data include but are not limited to wind speed, wind direction, temperature, humidity, light intensity, etc., which are obtained from on-site sensors or weather forecasts. Historical power data mainly refers to the power output of wind farms, in kilowatts (kW) or megawatts (MW). The correlation between meteorological data and power data forms the basis of power prediction, and the collected data will be used as input features of the model. Among them, the renewable energy power output records the actual power generation of a wind farm in kW or MW. In meteorological data such as wind speed, wind direction and temperature, the wind speed unit is m / s, the wind direction is expressed in degrees, and the temperature unit is Celsius.

[0028] S2: Preprocess the collected historical data. This includes removing outliers, processing missing values, and normalizing the data. By constructing a scatter plot of wind power renewable energy power output and meteorological factors, the distribution of abnormal data can be observed. Abnormal data includes but is not limited to records of low output power under high wind speed or high light conditions, or data with constant power and large fluctuations in meteorological conditions. These data are identified and eliminated. At the same time, all input features are normalized so that they are in the same numerical range to facilitate the stable training of subsequent models. In specific implementation, a scatter plot is first used to visualize the relationship between meteorological data and power output to determine abnormal points and abnormal patterns, and then all input features are normalized to ensure that different feature value ranges are similar and reduce the impact of numerical scale differences on model training.

[0029] S3: Construct a convolutional neural network model and extract short-term features from the data. The CNN model first extracts local features on meteorological data and historical power data through convolution operations to identify short-term dependencies in the power fluctuations of renewable energy. The convolution layer further automatically captures local feature patterns between meteorological data and power output, and extracts changing trends in time series data. Specifically, we can consider first constructing a multi-layer convolutional neural network, where each layer extracts short-term dependency features through convolution kernels, and then uses pooling layers to reduce feature dimensions, compressing the feature space while retaining important information and outputting local feature representations processed by the convolution layer as input for subsequent feature fusion.

[0030] S4: The local features extracted by the convolutional neural network are input into the Transformer model for global feature extraction. The Transformer captures the long-term dependencies between different time steps in the input sequence through the self-attention mechanism, and combines position encoding to ensure the model's ability to recognize time order information, effectively modeling the long-term time dependencies involved in power prediction. In particular, the multi-head self-attention mechanism is used to calculate the attention weights between each time step to identify long-term dependencies, and position encoding is used to add the position information in the time series to the model input, so that the model can capture the temporal features, and the final global feature representation is used as the output and combined with the local features for subsequent fusion.

[0031] S5: Perform multi-level feature fusion. Perform weighted summation or concatenation of local features extracted by the convolutional neural network and global features extracted by the Transformer to generate a multi-level fusion feature representation, thereby capturing local details and global trend information in the new energy power forecast and improving the forecast accuracy. Specifically, you can consider using a weighted summation method to add local features and global features weighted by importance; or use feature concatenation to retain the complete information of the two types of features, and then introduce the attention mechanism a second time to further optimize the correlation between different features, so that the final feature fusion result is more accurate.

[0032] S6: Perform multi-step prediction and output future power values. Based on the fused multi-level features, a fully connected layer is used to perform multi-step prediction of new energy power. By setting up multiple output nodes, power output prediction for multiple future time steps is achieved, ensuring that the prediction results cover a longer time range, thereby avoiding the accumulation of errors in step-by-step predictions. When constructing a fully connected layer with multiple output nodes, the fused features are input into the fully connected network, and the prediction results for each time step are output when the power for multiple future time steps is predicted simultaneously. In general, the prediction results need to be verified using actual data, and the accuracy of the predictions is improved by iteratively optimizing the model parameters.

[0033] Embodiment 3: Wind power prediction data model with dual-model fusion having a hierarchical architecture (Model A + Model B) Model A is a data model for wind power prediction that integrates a multi-level feature extraction algorithm of Transformer and convolutional neural network. On the basis of the above model A, an independent artificial intelligence model B is constructed to form a dual-model architecture. Among them, model B optimizes all adjustable parameters in model A (here refers to model A that has completed pre-training and iterative optimization).

[0034] Specifically, the core of model B is to build a special data model that integrates the following two elements: element x: meteorological spatiotemporal dynamic characteristics; element y: wind power generation scenario.

[0035] Factor x mainly considers the temporal and spatial order in which the temporal and spatial dynamic characteristics of meteorological conditions affect different factors y; for example, cold wave weather will affect wind power generation scenarios in different geographical locations in a certain temporal and spatial order. At this time, the temporal and spatial geographical dynamic characteristics of the current cold wave weather, namely factor x, and the different wind power generation scenarios it affects in succession are an ordered set, recorded as factor y=(yi, i=1, 2, 3, ...).

[0036] The working process of Model B is as follows: First, for the optimized A model, the B model pre-adjusts the adjustable parameters. The pre-adjustment generally adjusts the point value to a value range. Of course, various types of pre-adjustments can also be performed according to the specific situation. In any case, after the pre-adjustment, the A model actually corresponds to multiple models, which are recorded as the variant model set A′=(A′ j , j=1,2,3,…).

[0037] Then, each element in the mutation model set A′ will act on y=(y i , i=1, 2, 3, ...) in each element y i Because the cold wave weather in element x acts on each element y in time and space order, i Therefore, model B can perform post-hoc data observation, that is, observe each element A′ in the variant model set A′ j After acting on the first few elements of y (the first few elements of y are wind power generation scenarios where the cold wave has arrived), one or several variant models whose prediction effects are closest to the actual wind power data are identified, and a new set of variant models is obtained, denoted as A′′.

[0038] A′′ includes the optimized A model with better predictive efficiency in the current cold wave weather, so that the A model can be adjusted in real time based on A′′, so as to improve the predictive efficiency of the optimization model in real time and dynamically optimize the last few elements of factor y (the last few elements of y are wind power generation scenarios where the cold wave has not yet arrived).

[0039] Overall, the wind power data model under the two-layer data architecture, in terms of the fusion construction of model A, the method of the present invention realizes multi-level feature extraction of the time series of renewable energy power generation by integrating convolutional neural network and Transformer model, which can effectively capture local and global time-dependent features, thereby significantly improving the accuracy of power prediction. The present invention adopts a multi-step prediction strategy, and by setting multiple output nodes, it can simultaneously predict multiple future time steps, effectively avoiding the problem of error accumulation in step-by-step prediction, and improving the efficiency and timeliness of prediction. By introducing the self-attention mechanism and position encoding, the model can more flexibly capture long-range dependencies in complex time series, making up for the shortcomings of traditional machine learning methods in processing nonlinear and multidimensional time series data. The multi-level feature extraction algorithm based on CNN and Transformer can effectively improve the accuracy of renewable energy power prediction, solve the shortcomings of traditional prediction methods and single deep learning methods in modeling complex time series dependencies, and provide more accurate data support for the scheduling and planning of renewable energy grid-connected power generation.

[0040] At the same time, it is worth explaining that this fusion model is constructed based on wind power scenarios, but it can be seen that it is also applicable to processing multiple energy scenarios such as wind power and photovoltaics in new energy power generation, and has broad application prospects.

[0041] Furthermore, with regard to the construction of the two-layer data architecture of the A+B model, in the A model, the CNN convolution kernel sweeps across the spatial map of meteorological data, capturing spatial features of different scales, such as capturing micro-meteorological features such as local wind shear and temperature gradient; the Transformer multi-head attention mechanism cruises in the long river of time, keeping a close eye on the complex relationship between wind power and meteorology over time. The fusion of the two lays a solid characteristic foundation for wind power prediction, fully outlines the blueprint of the spatiotemporal impact of meteorology-power, makes the input model information richer and more accurate, and improves the prediction resolution from the root; on this basis, considering the orderly adaptation of dynamic and complex meteorological scenes, the B model is constructed to optimize the parameters of the A model, closely following the dynamic spatiotemporal characteristics of meteorology (such as the rhythm of cold waves invading wind farms in geographical order), and ensuring that the parameters of the A model keep pace with the times through pre-adjustment and practical screening of variant models, and flexibly responding to extreme and gradual meteorological changes, breaking the static shackles of traditional models, making the prediction model a "smart" model that dances with the weather, while enhancing the model's robustness and generalization in different wind power scenarios. In this way, in the real-time prediction of wind power, this dual-model architecture relies on real-time observation, evaluation, and screening by the B model to continuously inject optimization genes into the A model; whenever there is a weather fluctuation, it quickly adjusts its pace, accurately diagnoses the subsequent wind power output, and strangles the prediction error in the cradle, providing highly reliable and high-precision power forecasts for downstream links such as power grid dispatching and power market transactions, thus ensuring efficient grid connection of wind power and stable operation of the power system.

[0042] Example 4: Data-based example of the A+B model in application scenarios (taking cold wave weather events as an example) In the basic setting of the model architecture, the A model is considered to integrate the Transformer and convolutional neural network and use the convolution operation and multi-head attention mechanism calculation in mathematics to process the historical time series data of wind power and meteorological data. For the B model, from the perspective of the integration of meteorological spatiotemporal dynamic characteristics and wind power generation scenarios, it is necessary to construct data for the representation of meteorological elements, the calculation of meteorological transmission, and the pre-adjustment of the parameters of the A model. In terms of the data processing process, it is divided into an offline training stage and a real-time prediction and optimization stage. In the offline training stage, the A model is first trained, using the mean square error loss function and the back propagation algorithm, and the parameters are updated using stochastic gradient descent and its variants. Then the B model is constructed and initialized, the meteorological spatiotemporal dynamic characteristics and the power generation scenario sequence set are determined, and the A model parameters are pre-adjusted to generate a set of variant models. In the real-time prediction and optimization stage, the meteorological data is monitored in real time and the power generation scenarios are sorted. The variant model is applied to the power generation scenario and evaluated, the variant model is screened and optimized, and finally the updated model is used for real-time prediction, and the cycle continues until the meteorological event ends. The entire process combines deep learning model architecture design, meteorological physical process modeling and mathematical optimization methods, and is one of the core development projects of this technical team.

[0043] ①、A model related settings and processing Assume that the data input to model A includes the historical time series of wind power ( ) and the corresponding meteorological data (Contains a variety of meteorological information, recorded in time steps).

[0044] For weather data (The dimensions are assumed to be ), using convolutional neural network (CNN) to extract local spatial features. Through multiple convolution kernels (Size is , ) performs convolution operation to obtain feature map , the formula is: , where Conv represents the convolution operation, with a step size of Sliding convolution kernel calculation.

[0045] Wind power time series Concatenate with the meteorological features extracted by convolution into the input sequence X ( t ) = [ P ( t ) , F 1 ( t ) , F 2 ( t ) , ⋯ , F S ( t ) ] (The dimensions are adjusted to After that), it is input to the Transformer encoder.

[0046] Transformer processing and output: In the Transformer encoder, the Multi-HeadAttention mechanism is calculated as follows: Calculate first , , ( , , , is the learnable weight matrix).

[0047] Then ,in ( for dimensions).

[0048] at last ( for the number of heads).

[0049] The multi-head attention mechanism is followed by a feed-forward neural network (FFN) for nonlinear transformation to obtain the wind power prediction value output by model A. , the overall adjustable parameter set of model A is recorded as .

[0050] ②. B model related settings and processing Element x (meteorological spatiotemporal dynamic characteristics) represents: Gridding meteorological data in geographic space (divided into grid cells, , ). For meteorological elements (such as wind speed, wind direction, temperature, etc.) ,unit The meteorological feature vector is expressed as x i j ( t ) = [ v i j ( t ) , oh i j ( t ) , T i j ( t ) ] .

[0051] Consider the dynamics of meteorological propagation, such as the speed of cold wave propagation (two-dimensional vector), for the target wind farm location set ( is the number of power plants), the cold wave starts from the starting grid unit In time Start propagating and reach the power plant Time satisfy: (here represents the two-dimensional space distance measurement), sorting the power plants by arrival time to obtain an ordered set , corresponding elements (Wind power generation scenario).

[0052] B model core optimization process: For model A parameters Pre-adjust each parameter ( , is the total number of parameters) from point values ​​to value ranges, such as using normal distribution perturbations to generate parameter sets (corresponding to the variant model set A'), where each element satisfy: ,here ( , The number of variant models generated for pre-tuning, is the standard deviation controlling the disturbance amplitude, is the identity matrix).

[0053] ③. Data processing flow Offline training phase: A model training: Leveraging historical wind power data and weather data (Divided into training set, validation set, test set), using the mean square error (MSE) loss function: .

[0054] Update parameters using stochastic gradient descent (SGD) and its variants (such as the Adam optimizer) through the back-propagation algorithm (gradient calculation based on the chain rule, such as the derivation of convolutional layer weights involves the derivation rule of convolution operations, and the derivation of Transformer weights is performed module by module according to its structure) , train until convergence (validation set loss no longer decreases), and obtain the initial better A model.

[0055] B model construction and initialization: Based on the statistical analysis of historical meteorological data, the spatiotemporal dynamic characteristics of typical meteorological processes such as cold waves are determined, and the parameters such as the starting grid and propagation speed are constructed. Representation method, and organize the sequence sets corresponding to different power generation scenarios .

[0056] The parameters of the trained A model are adjusted in the above-mentioned way to generate the initial variant model set. When the cold wave and other meteorological events occur, the arrival time of each wind farm is calculated based on the meteorological spatiotemporal dynamic characteristic function, and the power generation scene is sorted and updated. .

[0057] Function and evaluation of mutation model: Each model in the variant model set A' Act on Elements (The corresponding meteorological and historical power data are input into each variation model to obtain the predicted power series ).

[0058] Using the rolling window evaluation method, calculate each variant model in the first few (set as ) The MSE error between the predicted power and the actual power in the power generation scenario (affected by the weather): .

[0059] Screening and optimizing mutation models: Based on calculation Sort the variant models and select the top ( ) mutation models form a new set A'', extract the corresponding A model parameter mean or weighted mean (weight is set according to the inverse ratio of error) as the optimized parameter , update model A.

[0060] Real-time prediction for wind power application scenarios: Using the updated model, combined with real-time meteorological and historical power data, the power of subsequent wind power generation scenarios (not fully affected by the weather) is predicted to achieve dynamic optimization of prediction efficiency, and the steps of the above real-time prediction and optimization stages are continuously cycled until the meteorological event ends.

[0061] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0062] In various embodiments, the hardware implementation of the technology can directly adopt existing intelligent devices, including but not limited to industrial computers, PCs, smart phones, handheld stand-alone computers, floor-standing stand-alone computers, etc. Its input device preferably adopts a screen keyboard, its data storage and calculation module adopts existing memory, calculator, controller, its internal communication module adopts existing communication ports and protocols, and its remote communication adopts existing GPRS network, World Wide Internet, etc.

[0063] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0064] In the embodiments provided by the present invention, it should be understood that the disclosed devices / terminal equipment and methods can be implemented in other ways. For example, the device / terminal equipment embodiments described above are only schematic, for example, the division of modules or units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms. The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the scheme of this embodiment.

[0065] Each functional unit in each embodiment of the present invention may be integrated into a processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of a software functional unit. If the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it may be stored in a computer-readable storage medium. Based on such an understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and may also instruct the relevant hardware to complete it through a computer program. The computer program may be stored in a computer-readable storage medium, and the computer program may be executed by the processor to implement the steps of the above-mentioned method embodiments. Among them, the computer program includes computer program code, and the computer program code may be in the form of source code, object code, executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc.

[0066] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A wind power prediction data model with hierarchical dual-model fusion is constructed based on the Transformer architecture and convolutional neural network to build a data layer, and the data layer is pre-trained using a pre-built annotated data set to obtain a wind power prediction data model with dual-model fusion, which is denoted as model A. It is characterized by: Furthermore, an independent data layer is constructed based on the A model, which is recorded as the B model, thereby forming a hierarchical data model architecture; wherein the A model serves as the lower-layer data model, the B model serves as the upper-layer data model, and the B model optimizes all adjustable parameters in the A model.

2. The hierarchical dual-model fusion wind power prediction data model according to claim 1 is characterized by: The B model is constructed as a data model integrating the following two data elements: element x: meteorological spatiotemporal dynamic characteristics; element y: wind power generation scenario.

3. The hierarchical dual-model fusion wind power prediction data model according to claim 2 is characterized by: The relationship between factor x and factor y is: factor x considers the spatiotemporal order of the impact of different factors y due to the spatiotemporal dynamic characteristics of meteorological conditions; factor x affects wind power generation scenarios in different geographical locations in sequence according to the spatiotemporal order, and the different wind power generation scenarios affected successively correspond to an ordered data set.

4. The hierarchical dual-model fusion wind power prediction data model according to claim 3 is characterized by: The B model is constructed as the upper structure of the A model. First, for the pre-trained A model, the B model pre-adjusts the adjustable parameters therein, and obtains the variant model set A′ after the pre-adjustment; then, each element in the variant model set A′ acts on each element in the ordered data set corresponding to the element y; since the element x acts on each element of the element y in sequence according to the time and space order, the B model can perform data observation, that is, after observing that each element in the variant model set A′ acts on the first few elements of y, one or several variant models whose prediction effect is closest to the actual wind power data are identified, and a new variant model set is obtained, denoted as A′′, which includes the optimized A model with better prediction efficiency in the current cold wave weather; in this way, the A model is adjusted in real time based on A′′, so as to realize the real-time dynamic optimization of wind power prediction for the last few elements in the ordered data set corresponding to the element y.

5. The hierarchical dual-model fusion wind power prediction data model according to claim 1 is characterized by: The model A integrates the multi-level feature extraction of Transformer and convolutional neural network to construct a process-based data structure to obtain a wind power prediction data model integrating the two models; the process-based data structure of model A includes: data process structure one, preprocessing the historical wind power data and related meteorological data and generating an input feature matrix; data process structure two, using convolutional neural network CNN to extract local features of the preprocessed input data, capturing the short-term dependency in the wind power time series and generating a preliminary feature representation; data process structure three, inputting the preliminary feature representation in data process structure two into the Transformer model, extracting long-term time dependency features through the self-attention mechanism and enhancing the global representation capability of the features; data process structure four, performing hierarchical feature fusion of the local features extracted by CNN and the global features extracted by Transformer to generate a comprehensive feature representation; data process structure five, performing multi-step prediction of wind power based on the fused features and using a fully connected layer to output the final prediction result.

6. The hierarchical dual-model fusion wind power prediction data model according to claim 5 is characterized by: The convolutional neural network in the data process structure 2 adopts a multi-layer convolution structure to abstract the local features in the time series layer by layer and use the pooling layer to reduce the feature dimension.

7. The hierarchical dual-model fusion wind power prediction data model according to claim 5 is characterized by: The Transformer model in data process structure three adopts a multi-layer encoder structure. Each layer of the encoder includes a self-attention mechanism and a feedforward neural network. The multi-head self-attention mechanism is used to capture the dependencies between different time steps in the input sequence, and the global information of the sequence is captured by calculating the correlation weights between each time step and other time steps. At the same time, the position information in the time series is explicitly encoded into the input features in combination with position encoding.

8. The hierarchical dual-model fusion wind power prediction data model according to claim 5 is characterized in that: The multi-level feature fusion in the data process structure 4 effectively fuses the local features extracted by CNN and the global features extracted by Transformer by weighted summation or splicing; CNN extracts short-term dependency features from the local window and captures the local patterns and detail features in the wind power time series, and Transformer can extract global information and model dependencies over a long period of time through the self-attention mechanism; the accuracy of the prediction is improved by fusing the features of the two and utilizing both short-term and long-term information; in the process of feature fusion, the weighted summation is used to dynamically adjust the weights according to the importance of local and global features, or feature splicing is used to retain more original information; in addition, an additional attention mechanism is introduced after the fusion to improve the model's sensitivity to wind power fluctuations and prediction accuracy by learning the correlation between different features (focusing on features that are more important to the prediction task).

9. The hierarchical dual-model fusion wind power prediction data model according to claim 5 is characterized by: The prediction part in the data process structure five adopts a multi-step prediction algorithm structure. By setting multiple output nodes, it simultaneously predicts the changes in wind power in multiple time steps in the future, avoids the problem of error accumulation in step-by-step prediction, and directly gives the power prediction results for multiple time points in the future within one prediction cycle; each output node corresponds to the predicted value of a specific time step in the future. The model learns the dependency between each time step in the time series through training and then makes a comprehensive prediction of the power trend in the future time period.