A method for predicting the output of new energy power plants

Abstract

This invention relates to the field of data reasoning technology and discloses a method for predicting the output of new energy power plants. The method includes: extracting environmental features from meteorological data and encoding them into environmental feature vectors; combining the inverter changes and component attenuation characteristics of the new energy power plant into an equipment feature vector; fusing the environmental feature vector and the equipment feature vector into a dynamic feature vector of the new energy power plant; predicting the initial output value of the new energy power plant based on the dynamic feature vector; constructing a spatial relationship matrix based on the latitude and longitude coordinates of the new energy power plant, superimposing the light intensity from the meteorological data into the spatial relationship matrix to obtain the spatial weight of the new energy power plant; when the new energy power plant is a hub power plant, strengthening the spatial weight based on the correlation of output fluctuations between the hub power plant and surrounding power plants, and predicting the final predicted value of the new energy power plant based on the strengthened spatial weight; this invention can improve the accuracy and speed of new energy power plant output prediction.

Description

Technical Field

[0001] This invention relates to the field of data reasoning technology, and in particular to a method for predicting the output of new energy power plants. Background Technology

[0002] In the field of power generation forecasting for renewable energy plants, existing technologies mainly rely on single-dimensional data features for modeling. For example, considering only meteorological data or equipment operation data makes it difficult to comprehensively capture the dynamic changes in renewable energy output. The lack of sufficient integration of the synergistic effects of environmental and equipment status characteristics results in insufficient adaptability of the forecasting models to complex scenarios, leading to low forecast accuracy.

[0003] Meanwhile, traditional forecasting methods lack effective modeling of spatial correlations, especially when dealing with the output fluctuation correlation between hub stations and surrounding stations. They cannot accurately construct spatial weight relationships, resulting in a failure to fully utilize the topological information of the regional power grid during the forecasting process. Furthermore, the lack of an iterative optimization mechanism prevents the model from adjusting its feature vectors in real time based on forecast deviations, further impacting forecast efficiency and accuracy, and making it difficult to meet the requirements of new energy grid-connected dispatch for forecasting rate and accuracy. Summary of the Invention

[0004] This invention provides a method for predicting the output of new energy power plants, the main purpose of which is to solve the problem of low accuracy and rate of power output prediction for new energy power plants.

[0005] To achieve the above objectives, the present invention provides a method for predicting the output of a new energy power station, comprising:

[0006] S1. Extract environmental features from meteorological data, encode them into environmental feature vectors, and combine the inverter changes and component attenuation features of new energy power plants into equipment feature vectors;

[0007] S2. The environmental feature vector and the equipment feature vector are fused into the dynamic feature vector of the new energy power station;

[0008] S3. Predict the initial power output of the new energy power station based on the dynamic feature vector;

[0009] S4. Construct a spatial relationship matrix based on the latitude and longitude coordinates of the new energy power station, and superimpose the light intensity from the meteorological data into the spatial relationship matrix to obtain the spatial weight of the new energy power station;

[0010] S5. When the new energy power station is a hub power station, the spatial weight is strengthened according to the correlation of the output fluctuation between the hub power station and the surrounding power stations, and the final predicted value of the new energy power station is predicted based on the strengthened spatial weight.

[0011] In a preferred embodiment, the step of extracting environmental features from meteorological data, encoding them into an environmental feature vector, and combining the inverter changes and component degradation characteristics of the new energy power station into an equipment feature vector includes:

[0012] Gaussian filtering was applied to the temperature, wind speed, and irradiance data in the meteorological data to obtain the environmental feature vector of the new energy power station.

[0013] The component power attenuation curves during historical operation are fitted to the attenuation rate parameters of the new energy power station;

[0014] The runtime is superimposed on the attenuation rate parameter to obtain the real-time attenuation coefficient of the energy station;

[0015] The real-time attenuation coefficient is combined with the inverter changes of the new energy power station to form the equipment feature vector of the energy power station.

[0016] In a preferred embodiment, fusing the environmental feature vector and the equipment feature vector into the dynamic feature vector of the new energy power station includes:

[0017] The covariance between wind speed, light intensity, and temperature in the environmental feature vector is encoded into an environmental covariance matrix.

[0018] The covariance between the inverter efficiency and the component attenuation coefficient of the device feature vector is encoded into a device covariance matrix.

[0019] The weight allocation ratio of the new energy power station is determined based on the eigenvalue distribution of the environmental covariance matrix and the equipment covariance matrix;

[0020] The environmental feature vector and the equipment feature are fused based on the weight allocation ratio to obtain the dynamic feature vector of the new energy power station.

[0021] In a preferred embodiment, predicting the initial power output of the renewable energy power station based on the dynamic feature vector includes:

[0022] Extract the temporal dependency features of the dynamic feature vector;

[0023] A linear transformation is performed on the time-dependent features to obtain the output probability distribution of the new energy power station;

[0024] The maximum probability in the output probability distribution is taken as the initial output value of the new energy power station.

[0025] In a preferred embodiment, constructing a spatial relationship matrix based on the latitude and longitude coordinates of the new energy power station includes:

[0026] The latitude and longitude coordinates of the new energy power station are converted into a set of position coordinates in a plane rectangular coordinate system;

[0027] A spatial relationship matrix is ​​generated based on the spherical distances of the new energy power stations in the location coordinate set.

[0028] In a preferred embodiment, the step of superimposing the light intensity from the meteorological data onto the spatial relationship matrix to obtain the spatial weights of the new energy power stations includes:

[0029] The spatial relationship matrix is ​​then scaled inversely.

[0030] The number of grid cells representing the light intensity is sampled to the location coordinates of the new energy power station to obtain the diagonal matrix of the light intensity of the new energy power station;

[0031] The light intensity diagonal matrix is ​​superimposed onto the spatial relationship matrix after inverse scaling to obtain the light enhancement matrix of the new energy power station.

[0032] Extract the main diagonal elements of the illumination enhancement matrix, and generate the spatial weights of the new energy power stations based on the main diagonal elements.

[0033] In a preferred embodiment, when the new energy power station is a hub power station, it includes:

[0034] When the grid-connected capacity of the new energy power station exceeds the regional threshold and it is located at a critical node of the power grid, it is marked as a hub power station;

[0035] Centered on the hub station, all new energy power stations within a preset radius are selected as associated power stations.

[0036] In a preferred embodiment, strengthening the spatial weight based on the output fluctuation correlation between the hub station and surrounding stations includes:

[0037] Based on the output sequence of the hub station and the associated stations, the fluctuation correlation between the hub station and the associated stations is calculated, wherein the calculation formula for the fluctuation correlation is as follows:

[0038] ;

[0039] In the formula, The aforementioned volatility correlation, For mathematical expectation, For the first The intensity of power output fluctuations at each hub station For the first The intensity of power output fluctuations at each associated power station For the first Output time-series data of each hub station For the first The intensity of power output fluctuations at each associated power station For the first The average output level of each hub station For the first The average power output level of each associated station;

[0040] The fluctuation correlation is arranged into a fluctuation matrix of the new energy power station according to the topological relationship between the hub station and the associated station;

[0041] The largest eigenvalue in the fluctuation matrix is ​​extracted as the feature vector corresponding to the new energy power station;

[0042] The spatial weights are enhanced based on the feature vectors to obtain enhanced spatial weights.

[0043] In a preferred embodiment, predicting the final value of the renewable energy power station based on the enhanced spatial weights includes:

[0044] The initial output value is divided into subsequences according to the time dimension;

[0045] The subsequence is convolved based on the enhanced spatial weights;

[0046] The final predicted value of the new energy power station is obtained by connecting the subsequence processed by residual convolution with the initial output value.

[0047] In a preferred embodiment, when the renewable energy power station is a hub power station, after strengthening the spatial weight based on the output fluctuation correlation between the hub power station and surrounding power stations, and predicting the final predicted value of the renewable energy power station based on the strengthened spatial weight, the process includes:

[0048] When the deviation between the initial output value and the final predicted value exceeds the preset tolerance, return to step S2 to update the dynamic feature vector;

[0049] The final predicted value of the new energy power station is re-predicted based on the updated dynamic feature vector.

[0050] Compared with the prior art, the present invention has the following beneficial effects:

[0051] 1. This invention constructs a dynamic feature vector by fusing environmental and equipment feature vectors, and generates spatial weights by combining light intensity and spatial relationship matrices from meteorological data. This comprehensively captures the influencing factors of new energy power plant output, thereby improving the accuracy of power plant output prediction. Simultaneously, it predicts initial output values ​​based on the dynamic feature vectors and strengthens spatial weights by leveraging the correlation between power plant output fluctuations and surrounding plants, achieving precise calculation of the final predicted value and effectively improving prediction accuracy.

[0052] 2. In the prediction process, when the deviation between the initial output value and the final predicted value exceeds the preset tolerance, the present invention can update the dynamic feature vector and predict again. This iterative optimization mechanism ensures the adaptability and accuracy of the model, further improves the accuracy and speed of power plant output prediction, and provides more reliable prediction support for new energy grid connection scheduling. Attached Figure Description

[0053] Figure 1 This is a flowchart illustrating a method for predicting the output of a new energy power station according to an embodiment of the present invention.

[0054] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0055] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0056] This application provides a method for predicting the power output of a renewable energy power plant. The execution entity of this method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, or other similar device. In other words, the method for predicting the power output of a renewable energy power plant can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0057] Reference Figure 1 The diagram shown is a flowchart illustrating a method for predicting the output of a new energy power station according to an embodiment of the present invention. In this embodiment, the method for predicting the output of a new energy power station includes:

[0058] S1. Extract environmental features from meteorological data, encode them into environmental feature vectors, and combine the inverter changes and component attenuation features of new energy power plants into equipment feature vectors;

[0059] In this embodiment of the invention, the step of extracting environmental features from meteorological data, encoding them into an environmental feature vector, and combining the inverter changes and component attenuation characteristics of new energy power plants into an equipment feature vector includes:

[0060] Gaussian filtering was applied to the temperature, wind speed, and irradiance data in the meteorological data to obtain the environmental feature vector of the new energy power station.

[0061] The component power attenuation curves during historical operation are fitted to the attenuation rate parameters of the new energy power station;

[0062] The runtime is superimposed on the attenuation rate parameter to obtain the real-time attenuation coefficient of the energy station;

[0063] The real-time attenuation coefficient is combined with the inverter changes of the new energy power station to form the equipment feature vector of the energy power station.

[0064] Specifically, temperature, wind speed, and radiation data are extracted from meteorological data, and Gaussian filtering is applied to these data.

[0065] Furthermore, Gaussian filtering is based on the characteristics of the Gaussian function to construct a smoothing filter. During data processing, a weighted average is calculated based on the values ​​of neighboring data points around each data point, according to the weights determined by the Gaussian function, thereby obtaining new and smoother temperature data, wind speed data, and irradiance data. These processed data are then combined to form the environmental feature vector of the new energy power station.

[0066] Furthermore, data on the change of component power over time during the historical operation of new energy power plants are collected. Using curve fitting methods, this data is fitted into a curve that reflects the degradation trend of component power. Relevant information describing the degree of component power degradation is extracted from this fitted curve. This information is the degradation rate parameter of the new energy power plant.

[0067] Furthermore, the current operating time of the new energy power station is obtained, and this operating time information is superimposed on the previously obtained attenuation rate parameter. Specifically, the influencing factors corresponding to the operating time are incorporated into the attenuation rate parameter to obtain a new value that reflects the power attenuation of the components at the current moment. This value is the real-time attenuation coefficient of the energy power station.

[0068] Furthermore, information on various changes in the inverters of the new energy power station during operation is obtained. The previously obtained real-time attenuation coefficient is combined with this inverter change information and organized into an ordered set. This set contains key information reflecting the operating status of the energy power station equipment, which is the equipment feature vector of the energy power station.

[0069] In summary, this invention extracts environmental features from meteorological data and encodes them into environmental feature vectors, which can comprehensively capture the impact of environmental factors such as temperature, wind speed, and radiation on the power output of new energy power plants. At the same time, Gaussian filtering is used to remove data noise, ensuring the accuracy and stability of environmental features and providing reliable environmental dimension data support for subsequent forecasting.

[0070] In summary, when the inverter's changes and the component's degradation characteristics are combined into a device feature vector, the degradation rate parameter is generated by fitting historical component power degradation curves, and the real-time degradation coefficient is obtained by superimposing the operating time. This can dynamically reflect the aging state of the equipment and the real-time operating efficiency, enabling the device feature vector to accurately characterize the hardware operating status of the new energy power station. This lays the foundation for constructing a dynamic feature vector that includes both environmental and equipment dimensions, effectively improving the adaptability and accuracy of the power output prediction model to dynamic changes in the equipment.

[0071] S2. The environmental feature vector and the equipment feature vector are fused into the dynamic feature vector of the new energy power station;

[0072] In this embodiment of the invention, fusing the environmental feature vector and the equipment feature vector into the dynamic feature vector of the new energy power station includes:

[0073] The covariance between wind speed, light intensity, and temperature in the environmental feature vector is encoded into an environmental covariance matrix.

[0074] The covariance between the inverter efficiency and the component attenuation coefficient of the device feature vector is encoded into a device covariance matrix.

[0075] The weight allocation ratio of the new energy power station is determined based on the eigenvalue distribution of the environmental covariance matrix and the equipment covariance matrix;

[0076] The environmental feature vector and the equipment feature are fused based on the weight allocation ratio to obtain the dynamic feature vector of the new energy power station.

[0077] Specifically, the environmental feature vectors that have been obtained include wind speed, light intensity, and temperature data.

[0078] Furthermore, covariance is used to measure the relationship between these data. By calculating the covariance between wind speed and light intensity, wind speed and temperature, and light intensity and temperature, and then arranging these covariance values ​​in a specific row and column order to form a matrix, this matrix is ​​the environmental covariance matrix, which reflects the degree of correlation between the data in the environmental feature vector.

[0079] Furthermore, the device feature vector contains data on inverter efficiency and component attenuation coefficients.

[0080] Furthermore, the same method of calculating covariance is used to measure the relationship between inverter efficiency and component attenuation coefficient. After calculating the covariance between the two, this covariance value is constructed into a matrix with only one element (or one value) according to the rules of the matrix, namely the device covariance matrix. This matrix reflects the correlation between the two data in the device feature vector.

[0081] Furthermore, after obtaining the environmental covariance matrix and the equipment covariance matrix, the distribution of eigenvalues ​​of these two matrices is analyzed.

[0082] Furthermore, eigenvalues ​​reflect the importance of the data features represented by the matrix. By identifying the larger eigenvalues ​​in the environmental covariance matrix that better reflect the important features of the environmental data, and the important eigenvalues ​​in the equipment covariance matrix, the importance of the environmental covariance matrix and the equipment covariance matrix in the whole can be determined based on the magnitude and distribution characteristics of these eigenvalues. This is then transformed into the weight allocation ratio of the new energy power station, that is, determining the proportion of environmental and equipment-related factors in the subsequent analysis.

[0083] Furthermore, after determining the weight allocation ratio, the environmental feature vector and the device feature vector are fused according to their respective weights.

[0084] Furthermore, the data in the environmental feature vector and the equipment feature vector are weighted separately, and then the weighted environmental feature vector data and the equipment feature vector data are added together to obtain a new set of data. The vector composed of this new set of data is the dynamic feature vector of the new energy power station. It comprehensively considers environmental and equipment factors and can reflect the dynamic operating status of the new energy power station.

[0085] In summary, when the present invention fuses environmental feature vectors and device feature vectors into a dynamic feature vector, it can accurately capture the dynamic correlation between environmental factors and device status by encoding the environmental covariance matrix and the device covariance matrix.

[0086] In summary, the weight allocation ratio is determined based on the eigenvalue distribution of the two types of covariance matrices. The fusion weights can be adaptively adjusted according to the importance of the features, so that the dynamic feature vector can reflect both the immediate impact of environmental changes on output and the attenuation effect of long-term conditions such as equipment aging on output.

[0087] In summary, the dynamic feature vector constructed by this fusion mechanism achieves the organic integration of environmental and device dimension data, forming a composite feature representation that includes spatiotemporal dynamic characteristics and device health status.

[0088] In summary, compared with single-dimensional feature modeling, this method can more comprehensively characterize the factors affecting the output of new energy power plants, provide richer feature inputs for subsequent initial output value prediction, effectively improve the adaptability of the prediction model to complex operating scenarios, and ensure the accuracy and dynamic response capability of output prediction from the feature fusion level.

[0089] S3. Predict the initial power output of the new energy power station based on the dynamic feature vector;

[0090] In this embodiment of the invention, predicting the initial power output of the renewable energy power station based on the dynamic feature vector includes:

[0091] Extract the temporal dependency features of the dynamic feature vector;

[0092] A linear transformation is performed on the time-dependent features to obtain the output probability distribution of the new energy power station;

[0093] The maximum probability in the output probability distribution is taken as the initial output value of the new energy power station.

[0094] Specifically, observe the dynamic feature vector, which contains environmental and equipment-related information of the new energy power station at different points in time.

[0095] Furthermore, by analyzing the patterns and interrelationships of these information over time, we can identify the temporal dependence characteristics present within them.

[0096] Furthermore, specifically, we examine the changing trends of data in the dynamic feature vectors of adjacent time points, as well as the influence relationships between data at different time points. We extract these features that reflect the temporal correlation to form a time-dependent feature set that can reflect the characteristics of the changes in new energy power station data over time.

[0097] Furthermore, the extracted temporal dependency features are processed using a linear transformation.

[0098] Furthermore, linear transformation operates on each data point in the time-dependent feature according to certain rules, thereby changing the relationships between the data.

[0099] Further, the specific operation involves assigning a specific weight to each time-dependent feature data, multiplying the data with the weights, and then summing all the product results to obtain a new set of data.

[0100] Furthermore, through this linear transformation, the time-dependent characteristics are transformed into a form that reflects the probability of different output conditions of new energy power plants, that is, the output probability distribution of new energy power plants is obtained. This distribution shows the probability of different output values ​​of new energy power plants in the future.

[0101] Furthermore, in the obtained power output probability distribution of the new energy power plants, we can examine the probability values ​​corresponding to each power output value.

[0102] Furthermore, the one with the highest probability value is identified, and the output value corresponding to this highest probability is determined as the initial output value of the new energy power station.

[0103] Furthermore, this initial output value is based on a comprehensive analysis of various factors such as the environment and equipment of the new energy power station, and is the most likely output scenario predicted, providing basic data for further analysis and prediction of the power output of the new energy power station.

[0104] In summary, this invention, when predicting the initial power output of renewable energy power plants based on dynamic feature vectors, captures the patterns of environmental and equipment state changes over time by extracting the temporal dependence features of the dynamic feature vectors, such as daily temperature variations and equipment degradation trends. A linear transformation is performed on the temporal dependence features to obtain the power output probability distribution, and the maximum probability value is used as the initial power output value. This probabilistic prediction method quantifies the possibilities of different power output scenarios, avoids the limitations of single-value prediction, and improves the reliability of the prediction results.

[0105] In summary, this prediction process is directly based on a dynamic feature vector that integrates environmental and equipment characteristics. It comprehensively reflects the real-time impact of fluctuating weather conditions and equipment operating status on power output. Compared to prediction methods that rely on only a single data dimension, it can more accurately depict the dynamic changes in power output at renewable energy power plants. Furthermore, the mechanism of determining the initial power output value through probability distribution ensures the statistical rationality of the prediction results, providing a more solid initial prediction foundation for subsequent optimization using spatial weights. This ensures the accuracy and scientific rigor of the initial power output prediction from the perspectives of time-series feature mining and probabilistic modeling.

[0106] S4. Construct a spatial relationship matrix based on the latitude and longitude coordinates of the new energy power station, and superimpose the light intensity from the meteorological data into the spatial relationship matrix to obtain the spatial weight of the new energy power station;

[0107] In this embodiment of the invention, constructing a spatial relationship matrix based on the latitude and longitude coordinates of the new energy power station includes:

[0108] The latitude and longitude coordinates of the new energy power station are converted into a set of position coordinates in a plane rectangular coordinate system;

[0109] A spatial relationship matrix is ​​generated based on the spherical distances of the new energy power stations in the location coordinate set.

[0110] The step of superimposing the light intensity from the meteorological data onto the spatial relationship matrix to obtain the spatial weights of the new energy power stations includes:

[0111] The spatial relationship matrix is ​​then scaled inversely.

[0112] The number of grid cells representing the light intensity is sampled to the location coordinates of the new energy power station to obtain the diagonal matrix of the light intensity of the new energy power station;

[0113] The light intensity diagonal matrix is ​​superimposed onto the spatial relationship matrix after inverse scaling to obtain the light enhancement matrix of the new energy power station.

[0114] Extract the main diagonal elements of the illumination enhancement matrix, and generate the spatial weights of the new energy power stations based on the main diagonal elements.

[0115] Specifically, the latitude and longitude coordinates of the new energy power stations are obtained, and these coordinates are converted into position coordinates in a Cartesian coordinate system using coordinate transformation methods. Specifically, a specific transformation method is used to map latitude and longitude points on the Earth's surface onto a plane, obtaining the corresponding x and y coordinate values. These plane coordinate values ​​of all new energy power stations are collected to form a set, which is the set of position coordinates in a Cartesian coordinate system.

[0116] Furthermore, based on the obtained set of location coordinates, the spherical distances between each renewable energy power station are calculated. Spherical distance refers to the shortest distance along a sphere between two points on the Earth's surface. After calculating the pairwise spherical distances between all renewable energy power stations, these distance values ​​are arranged into a matrix according to certain rules.

[0117] Furthermore, the rows and columns of the matrix correspond to different new energy power stations, and each element in the matrix represents the spherical distance between two new energy power stations. The resulting matrix is ​​a spatial relationship matrix, which shows the spatial positional relationship between various new energy power stations.

[0118] Specifically, take the previously obtained spatial relationship matrix and perform inverse scaling on each element in the matrix.

[0119] Furthermore, for each distance value in the spatial relationship matrix, a suitable value is found to divide it, such that the processed value decreases as the original distance value increases and increases as the original distance value decreases, thus obtaining a spatial relationship matrix after inverse scaling. This processing can highlight the relationship between new energy power stations that are relatively close.

[0120] Furthermore, since the light intensity is known to be distributed in a grid form, the grid data of the light intensity is obtained, and these grid data are sampled according to the position coordinates of the new energy power station in a Cartesian coordinate system.

[0121] Furthermore, the grid corresponding to the location coordinates of each new energy power station is located, and the light intensity value at that grid is extracted. Then, these light intensity values ​​are arranged in the order of the new energy power stations to form a matrix with non-zero elements only on the main diagonal and zero elements in all other positions. This matrix is ​​the light intensity diagonal matrix of the new energy power stations, which reflects the light intensity situation of each new energy power station.

[0122] Furthermore, the diagonal matrix of light intensity and the spatial relationship matrix after inverse scaling are superimposed. Specifically, the elements at corresponding positions in the two matrices are added to obtain a new matrix, which is the light enhancement matrix of the new energy power station. Through this superposition, light intensity information is integrated into the spatial relationship, highlighting the influence of light factors on the spatial relationship of the new energy power station.

[0123] Furthermore, observe the illumination enhancement matrix and extract its main diagonal elements, which are the elements on the line from the top left corner to the bottom right corner of the matrix.

[0124] Furthermore, these extracted diagonal elements are processed according to certain rules, such as sorting by numerical value and weighting by importance, to obtain a set of values ​​that reflect the spatial importance of the new energy power station. This set of values ​​is the spatial weight of the new energy power station, which comprehensively considers the impact of spatial distance and light intensity on the new energy power station.

[0125] In summary, when constructing a spatial relationship matrix based on the latitude and longitude coordinates of new energy power stations, this invention converts latitude and longitude into planar rectangular coordinates and generates a matrix based on spherical distance. This allows for an accurate characterization of the spatial location correlation between new energy power stations, laying a geometric foundation for subsequent analysis of the spatial correlation of power station output in the regional power grid.

[0126] In summary, this matrix construction method based on actual geographical coordinates can effectively reflect the transmission patterns of meteorological conditions between stations due to differences in geographical location, making the spatial relationship matrix physically reasonable.

[0127] In summary, when light intensity is superimposed onto the spatial relationship matrix, the weighting effect of nearby stations is enhanced by inverse scaling. By combining light intensity raster sampling to generate a diagonal matrix and superimposing it, the key environmental factor of light intensity can be deeply integrated with spatial location relationships, forming a spatial weight that includes both geographical distance and light intensity dimensions.

[0128] In summary, this weight not only reflects the spatial topological relationship between power stations, but also incorporates the direct impact of light intensity on power output. This allows the spatial weight to more comprehensively characterize the spatial correlation characteristics of power output from new energy power stations, providing more accurate weight support for strengthening spatial correlation calculations when predicting power output from hub power stations in the future. This, in turn, improves the accuracy of power output prediction and regional synergy from a spatial perspective.

[0129] S5. When the new energy power station is a hub power station, the spatial weight is strengthened according to the correlation of the output fluctuation between the hub power station and the surrounding power stations, and the final predicted value of the new energy power station is predicted based on the strengthened spatial weight.

[0130] In this embodiment of the invention, the step of when the new energy power station is a hub power station includes:

[0131] When the grid-connected capacity of the new energy power station exceeds the regional threshold and it is located at a critical node of the power grid, it is marked as a hub power station;

[0132] Centered on the hub station, all new energy power stations within a preset radius are selected as associated power stations.

[0133] The step of strengthening the spatial weight based on the output fluctuation correlation between the hub station and surrounding stations includes:

[0134] Based on the output sequence of the hub station and the associated stations, the fluctuation correlation between the hub station and the associated stations is calculated, wherein the calculation formula for the fluctuation correlation is as follows:

[0135] ;

[0136] In the formula, The aforementioned volatility correlation, For mathematical expectation, For the first The intensity of power output fluctuations at each hub station For the first The intensity of power output fluctuations at each associated power station For the first Output time-series data of each hub station For the first The intensity of power output fluctuations at each associated power station For the first The average output level of each hub station For the first The average power output level of each associated station;

[0137] The fluctuation correlation is arranged into a fluctuation matrix of the new energy power station according to the topological relationship between the hub station and the associated station;

[0138] The largest eigenvalue in the fluctuation matrix is ​​extracted as the feature vector corresponding to the new energy power station;

[0139] The spatial weights are enhanced based on the feature vectors to obtain enhanced spatial weights.

[0140] The prediction of the final value of the new energy power station based on the enhanced spatial weights includes:

[0141] The initial output value is divided into subsequences according to the time dimension;

[0142] The subsequence is convolved based on the enhanced spatial weights;

[0143] The final predicted value of the new energy power station is obtained by connecting the subsequence processed by residual convolution with the initial output value.

[0144] When the new energy power station is a hub power station, after strengthening the spatial weight based on the output fluctuation correlation between the hub power station and surrounding power stations, and predicting the final predicted value of the new energy power station based on the strengthened spatial weight, the process includes:

[0145] When the deviation between the initial output value and the final predicted value exceeds the preset tolerance, return to step S2 to update the dynamic feature vector;

[0146] The final predicted value of the new energy power station is re-predicted based on the updated dynamic feature vector.

[0147] Specifically, the grid-connected capacity data of new energy power plants and their node information in the power grid are obtained, while the regional thresholds and the specific range of key nodes in the power grid are determined.

[0148] Furthermore, the grid-connected capacity of each new energy power station is compared with the regional threshold. If the grid-connected capacity of a certain new energy power station is greater than the regional threshold, it is then checked whether the power station is within the range of a pre-determined key node of the power grid.

[0149] Furthermore, when the grid-connected capacity of a new energy power station exceeds the regional threshold and is located at a critical node of the power grid, the station is marked with a specific designation as a hub station, thereby distinguishing new energy power stations that hold an important position in the power grid.

[0150] Furthermore, after determining the hub station, the specific value of the preset radius is determined.

[0151] Furthermore, taking each hub station as the center, a circle with a preset radius is drawn on the map or power grid layout, and all new energy power stations within this circle are screened out. Regardless of the size or type of these power stations, as long as they are within this circular area, they are selected. These selected new energy power stations are collectively regarded as associated power stations, thereby determining the set of new energy power stations that have close spatial connections with the hub station.

[0152] Specifically, power output data from hub stations and related stations at different points in time are collected and organized in chronological order.

[0153] Furthermore, by analyzing the changes in power output data of the hub station and each associated station over time, the similarity of their power output fluctuations can be determined.

[0154] Furthermore, specifically, by comparing the timing and magnitude of the rise or fall in output between the two, the fluctuation correlation between the hub station and each associated station is calculated, resulting in a series of values ​​reflecting the degree of their fluctuation correlation.

[0155] Furthermore, the topological relationship between hub stations and associated stations should be clarified, that is, their connection method and layout structure in the power grid.

[0156] Furthermore, based on this topological relationship, the previously calculated fluctuation correlation values ​​are arranged.

[0157] Furthermore, the fluctuation correlation values ​​associated with each power station are filled into a matrix according to the connection order and correspondence of the power stations in the topology, so that the rows and columns of the matrix correspond to different power stations, and the elements in the matrix are the fluctuation correlations between two corresponding power stations. The matrix formed in this way is the fluctuation matrix of new energy power stations.

[0158] Furthermore, the fluctuation matrix is ​​analyzed to identify the largest eigenvalue.

[0159] Furthermore, eigenvalues ​​reflect the importance of the data features represented by the matrix. In the fluctuation matrix, the eigenvector corresponding to the largest eigenvalue contains the most important information about the power output fluctuations of the renewable energy power plants. Using a specific method, the data corresponding to the largest eigenvalue is extracted from the fluctuation matrix; this data is the eigenvector corresponding to the renewable energy power plant.

[0160] Furthermore, taking the previously obtained spatial weights and the newly generated feature vectors, the information in the feature vectors is incorporated into the spatial weights.

[0161] Furthermore, based on the important characteristics of the power output fluctuation of new energy power plants reflected by the feature vector, the values ​​of each in the spatial weight are adjusted.

[0162] For example, if the eigenvector shows that the fluctuations of certain stations have a significant impact on the overall situation, the corresponding values ​​of these stations in the spatial weight are increased accordingly. In this way, the spatial weight is strengthened, and the strengthened spatial weight is obtained, which can more accurately reflect the importance of new energy stations under comprehensive factors.

[0163] Specifically, the first step in the formula for calculating the volatility correlation... The output time-series data of each hub station is obtained by collecting the actual output values ​​of the hub station at different time points, and these values ​​arranged in chronological order are used as the output time-series data of that hub station; The output fluctuation intensity of each associated power station is determined by recording the output values ​​of the associated power stations at various time points, analyzing the changes of these values ​​over time, and thus deriving the degree of output fluctuation; The average output level of each hub station is obtained by summing all the output time-series data of that hub station and then dividing by the total number of data points; The average output level of each associated power station is calculated by summing all the output data of that associated power station and dividing by the number of data points; The intensity of power output fluctuations at each hub station is determined by observing the degree to which the station's power output time-series data deviates from the average power output level; the greater the deviation, the greater the fluctuation intensity. The intensity of power output fluctuation of a related power station is an indicator of the dispersion of the power output data of that related power station. It is obtained by calculating the square of the difference between each power output data and the average power output level, averaging these squared values, and then taking the square root.

[0164] Furthermore, this formula is used to calculate the correlation of fluctuations between hub stations and associated stations. By analyzing the deviation of the output data of the two stations from their respective average output levels, as well as the degree of dispersion of the output data, the degree of correlation of their output fluctuations can be determined.

[0165] Specifically, the method involves first calculating the product of the difference between the output data of each of the two power stations and the average output level, then taking the expected value of these products, and finally dividing them by a measure of the dispersion of the output data of each of the two power stations to obtain a value that reflects the correlation between the fluctuations of the two.

[0166] Furthermore, when the output fluctuations of the hub station and the associated station show similar trends, that is, when their output data deviate from the average output level in a similar way and the degree of dispersion is also similar, the calculated fluctuation correlation value will be large, indicating that the fluctuation correlation between the two is strong. If the output fluctuations of the two stations are not significantly related, when the output of one station increases while the output of the other station decreases, or when the degree of dispersion of their output data is very different, then the calculated fluctuation correlation value will be small, meaning that the fluctuation correlation between the two is weak.

[0167] Specifically, the initial power output data of the new energy power plants is obtained, and this data is recorded at different points in time.

[0168] Furthermore, these initial output values ​​are divided according to chronological order, grouping initial output values ​​within a continuous period of time into a subsequence. In this way, the entire initial output value data is split into multiple subsequences along the time dimension, and each subsequence contains output information within a specific time period.

[0169] Furthermore, the enhanced spatial weights obtained are used to perform convolution processing on the divided subsequences.

[0170] Furthermore, the specific operation involves treating the enhanced spatial weights as a "filter" and using this "filter" to perform a sliding operation on each subsequence.

[0171] Furthermore, during the sliding process, the data of the subsequence covered by the "filter" is multiplied by the corresponding weight value of the "filter" itself, and then these products are added together to obtain a new value. As the "filter" slides sequentially on the subsequence, a new value is obtained with each slide. These new values ​​are combined to form the subsequence after convolution processing. In this way, the information contained in the spatial weights is incorporated into the subsequence data.

[0172] Furthermore, a residual concatenation operation is performed between the convolutional subsequence and the original initial output value.

[0173] Furthermore, the data at each time point in the convolutionally processed subsequence is added to the corresponding time point data in the initial output value to obtain a new set of data. This new set of data integrates the features of the original data and the data after convolution processing, and can more comprehensively reflect the changes in the output of the renewable energy power station. This new set of data is used as the final predicted value for the renewable energy power station to predict its future output.

[0174] In summary, when a new energy power station serves as a hub station, this invention strengthens spatial weight by exploring the correlation between its output fluctuations and those of surrounding power stations, thereby accurately capturing the coordinated change patterns of power station output in the regional power grid.

[0175] In summary, the system identifies hub stations based on grid-connected capacity and grid node locations, selects associated stations centered on these hubs, quantifies the degree of coordination of power output fluctuations among stations using a fluctuation correlation formula, and strengthens spatial weights by transforming this correlation into eigenvectors of the fluctuation matrix. This ensures that spatial weights not only include geographical distance and light intensity information but also incorporate dynamic power output correlation characteristics, thereby constructing a spatial correlation model that better reflects actual operating scenarios and providing more scientific weight support for final prediction.

[0176] In summary, this enhancement mechanism achieves in-depth modeling of the power output fluctuation patterns around hub stations through the dual constraints of topological relationships and fluctuation correlations, enabling the prediction process to fully utilize the spatial coupling characteristics of the regional power grid.

[0177] In summary, when predicting the final value, by dividing the initial output value into subsequences and performing convolution processing with enhanced spatial weights, and then fusing the initial prediction with spatial correlation features through residual connections, the prediction model's ability to capture the output fluctuations of hub stations can be effectively improved. In particular, under complex meteorological conditions, it can more accurately reflect the coordinated change trend of regional output, thereby significantly improving the accuracy of new energy station output prediction and the efficiency of regional coordinated prediction.

[0178] In the several embodiments provided by this invention, it should be understood that the disclosed method can be implemented in other ways.

[0179] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0180] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, and technology that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0181] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for predicting the output of a new energy power station, characterized in that, The method includes: S1. Extract environmental features from meteorological data, encode them into environmental feature vectors, and combine the inverter changes and component attenuation features of new energy power plants into equipment feature vectors; S2. The environmental feature vector and the equipment feature vector are fused into the dynamic feature vector of the new energy power station; S3. Predict the initial power output of the new energy power station based on the dynamic feature vector; S4. Construct a spatial relationship matrix based on the latitude and longitude coordinates of the new energy power station, and superimpose the light intensity from the meteorological data into the spatial relationship matrix to obtain the spatial weight of the new energy power station; S5. When the new energy power station is a hub station, the spatial weight is strengthened based on the correlation between the output fluctuations of the hub station and surrounding stations, including: When the grid-connected capacity of the new energy power station exceeds the regional threshold and it is located at a critical node of the power grid, it is marked as a hub power station; Centered on the hub station, all new energy power stations within a preset radius are selected as associated power stations; Based on the output sequence of the hub station and the associated stations, the fluctuation correlation between the hub station and the associated stations is calculated, wherein the calculation formula for the fluctuation correlation is as follows: ； In the formula, The aforementioned volatility correlation, For mathematical expectation, For the first The intensity of power output fluctuations at each hub station For the first The intensity of power output fluctuations at each associated power station For the first Output time-series data of each hub station For the first Output time-series data of each associated power station For the first The average output level of each hub station For the first The average power output level of each associated station; The fluctuation correlation is arranged into a fluctuation matrix of the new energy power station according to the topological relationship between the hub station and the associated station; The largest eigenvalue in the fluctuation matrix is ​​extracted as the feature vector corresponding to the new energy power station; The spatial weights are enhanced based on the feature vectors to obtain enhanced spatial weights. The final predicted value of the new energy power station is predicted based on the enhanced spatial weights.

2. The power output prediction method for new energy power plants as described in claim 1, characterized in that, The environmental features extracted from meteorological data are encoded into environmental feature vectors. The inverter variations and component degradation characteristics of new energy power plants are combined into equipment feature vectors, including: Gaussian filtering was applied to the temperature, wind speed, and irradiance data in the meteorological data to obtain the environmental feature vector of the new energy power station. The component power attenuation curves during historical operation are fitted to the attenuation rate parameters of the new energy power station; The runtime is superimposed on the attenuation rate parameter to obtain the real-time attenuation coefficient of the energy station; The real-time attenuation coefficient is combined with the inverter changes of the new energy power station to form the equipment feature vector of the energy power station.

3. The method for predicting the output of new energy power plants as described in claim 1, characterized in that, The process of fusing the environmental feature vector and the equipment feature vector into the dynamic feature vector of the new energy power station includes: The covariance between wind speed, light intensity, and temperature in the environmental feature vector is encoded into an environmental covariance matrix. The covariance between the inverter efficiency and the component attenuation coefficient of the device feature vector is encoded into a device covariance matrix. The weight allocation ratio of the new energy power station is determined based on the eigenvalue distribution of the environmental covariance matrix and the equipment covariance matrix; The environmental feature vector and the equipment feature are fused based on the weight allocation ratio to obtain the dynamic feature vector of the new energy power station.

4. The method for predicting the output of new energy power plants as described in claim 1, characterized in that, The prediction of the initial power output of the renewable energy power station based on the dynamic feature vector includes: Extract the temporal dependency features of the dynamic feature vector; A linear transformation is performed on the time-dependent features to obtain the output probability distribution of the new energy power station; The maximum probability in the output probability distribution is taken as the initial output value of the new energy power station.

5. The method for predicting the output of new energy power plants as described in claim 1, characterized in that, The construction of the spatial relationship matrix based on the latitude and longitude coordinates of the new energy power station includes: The latitude and longitude coordinates of the new energy power station are converted into a set of position coordinates in a plane rectangular coordinate system; A spatial relationship matrix is ​​generated based on the spherical distances of the new energy power stations in the location coordinate set.

6. The method for predicting the output of new energy power plants as described in claim 1, characterized in that, The step of superimposing the light intensity from the meteorological data onto the spatial relationship matrix to obtain the spatial weights of the new energy power stations includes: The spatial relationship matrix is ​​then scaled inversely. The number of grid cells representing the light intensity is sampled to the location coordinates of the new energy power station to obtain the diagonal matrix of the light intensity of the new energy power station; The light intensity diagonal matrix is ​​superimposed onto the spatial relationship matrix after inverse scaling to obtain the light enhancement matrix of the new energy power station. Extract the main diagonal elements of the illumination enhancement matrix, and generate the spatial weights of the new energy power stations based on the main diagonal elements.

7. The method for predicting the output of new energy power plants as described in claim 1, characterized in that, The prediction of the final value of the new energy power station based on the enhanced spatial weights includes: The initial output value is divided into subsequences according to the time dimension; The subsequence is convolved based on the enhanced spatial weights; The final predicted value of the new energy power station is obtained by connecting the subsequence processed by residual convolution with the initial output value.

8. The method for predicting the output of new energy power plants as described in claim 1, characterized in that, When the new energy power station is a hub power station, after strengthening the spatial weight based on the output fluctuation correlation between the hub power station and surrounding power stations, and predicting the final predicted value of the new energy power station based on the strengthened spatial weight, the process includes: When the deviation between the initial output value and the final predicted value exceeds the preset tolerance, return to step S2 to update the dynamic feature vector; The final predicted value of the new energy power station is re-predicted based on the updated dynamic feature vector.

Images