Power system net load power fluctuation self-adaptive online generation method

Abstract

The invention discloses a power system net load power fluctuation self-adaptive online generation method and system, and the method comprises the steps: carrying out the feature extraction based on the historical operation data of a power grid and the weather data of a place where the power grid is located, and obtaining the feature data; performing feature screening by taking the wind power fluctuation power, the photovoltaic fluctuation power and the load fluctuation power as target variables; determining an optimal clustering number based on the target features, carrying out clustering based on the optimal clustering number, and respectively determining a clustering group corresponding to each target variable; obtaining a real-time feature sample, calculating the Euclidean distance between the real-time feature sample and the clustering center of each clustering group, and obtaining a target clustering group; calculating a power fluctuation interval of the real-time feature sample under a preset confidence coefficient based on the statistical features of the target variables in each target clustering group so as to determine a confidence interval; and determining a confidence interval of the net load fluctuation power based on the net load fluctuation power probability distribution, and automatically generating the net load fluctuation power based on the confidence interval.

Description

Technical Field

[0001] The present invention relates to the technical field of power system operation and control, and more specifically, to a method and system for adaptively and online generating power system net load power fluctuations. Background Art

[0002] The operation of a high-proportion new energy power system faces severe challenges. On the one hand, the demand on the load side shows highly time-varying characteristics, and its fluctuation amplitude and peak-valley difference increase significantly with the large-scale access of new types of electrical equipment. On the other hand, the power generation of new energy sources such as wind power and photovoltaic power is strongly coupled by meteorological conditions, and the output shows randomness and intermittency characteristics of multi-time-scale coupling. The superposition of uncertainties on both the source and load sides leads to an exponential increase in the complexity of system power balance. Under this background, the effective generation of the active power fluctuation range has become the core basis for ensuring the safe and economic operation of the power grid. The dispatching agency urgently needs to quantitatively evaluate the power disturbance boundaries at different time scales (from seconds to hours) to provide safety constraint boundaries for day-ahead and real-time multi-stage dispatching.

[0003] The existing method for constructing a static fault set based on deterministic scenarios is difficult to characterize the continuous disturbance characteristics of a new energy-dominated system. Therefore, there is an urgent need to develop a dynamic disturbance scenario generation technology that integrates meteorological-time-power correlation characteristics to provide theoretical support for the panoramic security defense system of a new power system.

[0004] Therefore, a method for adaptively and online generating power system net load power fluctuations is needed. Summary of the Invention

[0005] The present invention proposes a method and system for adaptively and online generating power system net load power fluctuations to solve the problem of how to online generate the comprehensive fluctuation range of active power that can support the safe operation of the power grid, and is applicable to regional or provincial synchronous power grids with a new energy penetration rate exceeding 30%.

[0006] To solve the above problems, according to one aspect of the present invention, a method for adaptively and online generating power system net load power fluctuations is provided, characterized in that the method includes:

[0007] Performing feature extraction based on the historical operation data of wind power generation power, photovoltaic power generation power, and load power with time stamps in the power grid and the weather data of the location of the power grid to obtain feature data;

[0008] Respectively using the wind power fluctuation power, photovoltaic power fluctuation power, and load fluctuation power as target variables for feature screening, and screening out target features that have a great influence on the wind power fluctuation power, photovoltaic power fluctuation power, and load fluctuation power;

[0009] Determine the optimal number of clusters corresponding to different target variables based on the target features corresponding to the different target variables, and perform clustering based on the optimal number of clusters to respectively determine the clustering groups corresponding to each target variable;

[0010] Extract features based on the real-time operation data and weather data of the power grid to obtain real-time feature samples, calculate the Euclidean distances between the real-time feature samples and the cluster centers of each clustering group, and determine the target clustering groups corresponding to each target variable based on the nearest Euclidean distance;

[0011] Calculate the statistical features of the target variables corresponding to each target clustering group respectively, and calculate the probability distribution of the net load fluctuation power based on the statistical features;

[0012] Determine the confidence interval of the net load fluctuation power based on the probability distribution of the net load fluctuation power, and automatically generate the net load fluctuation power based on the confidence interval.

[0013] Preferably, the feature data includes three categories: power fluctuation features, meteorological features, and time features;

[0014] Among them, the power fluctuation features include: wind power active power, photovoltaic active power, load active power, wind power fluctuation power, photovoltaic fluctuation power, and load fluctuation power; the meteorological features include: temperature, wind speed, and irradiance; the time features include: month, hour, day of the week, whether it is a weekend, and whether it is a holiday.

[0015] Preferably, the method further includes:

[0016] Before performing feature screening with wind power fluctuation power, photovoltaic fluctuation power, and load fluctuation power as target variables respectively, perform standardization processing on the feature data.

[0017] Preferably, when performing feature screening with wind power fluctuation power, photovoltaic fluctuation power, and load fluctuation power as target variables respectively, and screening out the target features that have a great impact on the power fluctuation power, photovoltaic fluctuation power, and load fluctuation power, it includes:

[0018] Screen out the features that have a great impact on the wind power fluctuation power based on the tree model as the target features;

[0019] Screen out the features that have a great impact on the photovoltaic fluctuation power based on PCA principal component analysis as the target features;

[0020] Screen out the features that have a great impact on the load fluctuation power based on the tree model recursive feature elimination as the target features.

[0021] Preferably, the determining the optimal number of clusters corresponding to the same target variable based on the target features corresponding to different target variables includes:

[0022] Draw the SSE curves for the target features corresponding to the wind power fluctuation power, photovoltaic power fluctuation power, and load power fluctuation power as target variables, and use the elbow method for calculation to respectively determine the optimal number of clusters corresponding to the wind power fluctuation power, photovoltaic power fluctuation power, and load power fluctuation power.

[0023] Preferably, the method further includes:

[0024] Perform standardization processing on the real-time feature samples.

[0025] Preferably, calculating the probability distribution of the net load fluctuation power based on the statistical features includes:

[0026] Construct a covariance matrix and a joint probability distribution respectively based on the means and variances of the target variables corresponding to the target clustering groups corresponding to the wind power fluctuation power, photovoltaic power fluctuation power, and load power fluctuation power; wherein, the covariance matrix is:

[0027]

[0028] The joint probability distribution is:

[0029]

[0030] Determining the probability distribution of the net load fluctuation power based on the variance matrix and the joint probability distribution includes:

[0031]

[0032] where, ΔP_NetLoad is the net load fluctuation power; and are respectively the mean and standard deviation of the wind power fluctuation power corresponding to the target clustering group of the wind power fluctuation power; and are respectively the mean and standard deviation of the photovoltaic power fluctuation power corresponding to the target clustering group of the photovoltaic power fluctuation power; and are respectively the mean and standard deviation of the photovoltaic power fluctuation power corresponding to the target clustering group of the load power fluctuation power; CM is the covariance matrix; ΔP_Wind, ΔP_Solar, and ΔP_Load are respectively the active power of wind charge fluctuation, active power of photovoltaic fluctuation, and load power fluctuation; cov() represents the correlation coefficient calculation function; μ net and σ net are respectively the mean and standard deviation of the net load fluctuation power.

[0033] Preferably, determining the confidence interval of the net load fluctuation power based on the probability distribution of the net load fluctuation power includes:

[0034] Determine a preset confidence level, and convert the preset confidence level into a Z value according to the standard normal distribution table;

[0035] Determine that the confidence interval of the net load fluctuation power at the preset confidence level is: [μ net -Z·σ net , μ net +Z·σ net ; where μ net and σ net are respectively the mean and standard deviation of the net load fluctuation power; Z represents the Z value corresponding to the preset confidence level.

[0036] According to another aspect of the present invention, there is provided an adaptive online generation system for the net load power fluctuation of a power system, and the system includes:

[0037] A feature data acquisition unit, configured to perform feature extraction based on the historical operation data of the wind power generation power, photovoltaic power generation power, and load power with time stamps in the power grid and the weather data of the location of the power grid, and acquire feature data;

[0038] A feature screening unit, configured to perform feature screening with the wind power fluctuation power, photovoltaic power fluctuation power, and load power fluctuation power as target variables respectively, and screen out the target features that have a great influence on the wind power fluctuation power, photovoltaic power fluctuation power, and load power fluctuation power;

[0039] A clustering unit, configured to determine the optimal number of clusters corresponding to different target variables based on the target features corresponding to different target variables, and perform clustering based on the optimal number of clusters to respectively determine the clustering groups corresponding to each target variable;

[0040] A target clustering group determination unit, configured to perform feature extraction based on the real-time operation data and weather data of the power grid to obtain real-time feature samples, calculate the Euclidean distance between the real-time feature samples and the cluster centers of each clustering group, and determine the target clustering group corresponding to each target variable based on the nearest Euclidean distance;

[0041] A probability distribution determination unit, configured to calculate the statistical features of the target variables corresponding to each target clustering group respectively, and calculate the net load fluctuation power probability distribution based on the statistical features;

[0042] A confidence interval determination unit, configured to determine the confidence interval of the net load fluctuation power based on the net load fluctuation power probability distribution, and perform automatic generation of the net load fluctuation power based on the confidence interval.

[0043] Preferably, the feature data includes three categories: power fluctuation features, meteorological features, and time features;

[0044] Among them, the power fluctuation characteristics include: wind power active power, photovoltaic power active power, load active power, wind power fluctuation power, photovoltaic power fluctuation power, and load fluctuation power; the meteorological characteristics include: temperature, wind speed, and irradiance; the time characteristics include: month, hour, day of the week, whether it is a weekend, and whether it is a holiday.

[0045] Preferably, the system further includes:

[0046] A normalization processing unit, configured to perform normalization processing on the feature data before performing feature screening with the wind power fluctuation power, photovoltaic power fluctuation power, and load fluctuation power as target variables respectively.

[0047] Preferably, the feature screening unit performs feature screening with the wind power fluctuation power, photovoltaic power fluctuation power, and load fluctuation power as target variables respectively, and screens out the target features that have a great impact on the electric power fluctuation power, photovoltaic power fluctuation power, and load fluctuation power, including:

[0048] Screen out the features that have a great impact on the wind power fluctuation power based on the tree model as the target features;

[0049] Screen out the features that have a great impact on the photovoltaic power fluctuation power based on PCA principal component analysis as the target features;

[0050] Screen out the features that have a great impact on the load fluctuation power based on the tree model recursive feature elimination as the target features.

[0051] Preferably, the clustering unit determines the optimal number of clusters corresponding to the same target variable based on the target features corresponding to different target variables, including:

[0052] Respectively draw SSE curves based on the target features corresponding to the wind power fluctuation power, photovoltaic power fluctuation power, and load fluctuation power as target variables, and use the elbow method for calculation to respectively determine the optimal number of clusters corresponding to the wind power fluctuation power, photovoltaic power fluctuation power, and load fluctuation power.

[0053] Preferably, the system further includes:

[0054] A normalization processing unit, configured to perform normalization processing on the real-time feature samples.

[0055] Preferably, the probability distribution determination unit calculates the probability distribution of the net load fluctuation power based on the statistical features, including:

[0056] Respectively construct a covariance matrix and a joint probability distribution based on the means and variances of the target variables corresponding to the target clustering groups corresponding to the wind power fluctuation power, photovoltaic power fluctuation power, and load fluctuation power; among them, the covariance matrix is:

[0057]

[0058] The joint probability distribution is as follows:

[0059]

[0060] Determining the probability distribution of the net load fluctuation power based on the variance matrix and the joint probability distribution includes:

[0061]

[0062] where ΔP_NetLoad is the net load fluctuation power; and are the mean and standard deviation of the wind power fluctuation corresponding to the target clustering group of wind power fluctuations, respectively; and are the mean and standard deviation of the photovoltaic power fluctuation corresponding to the target clustering group of photovoltaic power fluctuations, respectively; and are the mean and standard deviation of the photovoltaic power fluctuation corresponding to the target clustering group of load fluctuations, respectively; CM is the covariance matrix; ΔP_Wind, ΔP_Solar, and ΔP_Load are the active power fluctuations of wind power, photovoltaic power, and load fluctuations, respectively; cov() represents the correlation coefficient calculation function; μ net and σ net are the mean and standard deviation of the net load fluctuation power, respectively.

[0063] Preferably, the confidence interval determination unit determines the confidence interval of the net load fluctuation power based on the probability distribution of the net load fluctuation power, including:

[0064] Determine a preset confidence level and convert the preset confidence level to a Z value according to the standard normal distribution table;

[0065] Determine that the confidence interval of the net load fluctuation power at the preset confidence level is: [μ net -Z·σ net , μ net +Z·σ net ; where μ net and σ net are the mean and standard deviation of the net load fluctuation power, respectively; Z represents the Z value corresponding to the preset confidence level.

[0066] On the other hand of the present invention, the present invention provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, it implements the steps of any one of the methods for adaptively and online generating the net load power fluctuation of a power system.

[0067] According to another aspect of the present invention, the present invention provides an electronic device, including:

[0068] The computer-readable storage medium mentioned above; and one or more processors configured to execute the program in the computer-readable storage medium.

[0069] The present invention provides a method and system for adaptive online generation of net load power fluctuations in an electric power system, comprising: extracting features based on historical operating data of wind power generation power, photovoltaic power generation power, and load power with time stamps of a power grid and weather data at the location of the power grid to obtain feature data; performing feature screening using wind power fluctuation power, photovoltaic power fluctuation power, and load fluctuation power as target variables to screen out target features that have a significant impact on wind power fluctuation power, photovoltaic power fluctuation power, and load fluctuation power; determining the optimal number of clusters corresponding to different target variables based on the target features corresponding to different target variables, and performing clustering based on the optimal number of clusters. Determine the cluster group corresponding to each target variable respectively; perform feature extraction based on the real-time operation data and weather data of the power grid, obtain real-time feature samples, calculate the Euclidean distance between the real-time feature samples and the cluster center of each cluster group, and determine the target cluster group corresponding to each target variable based on the nearest Euclidean distance; calculate the statistical characteristics of the target variables corresponding to each target cluster group respectively, and calculate the net load fluctuation power probability distribution based on the statistical characteristics; determine the confidence interval of the net load fluctuation power based on the net load fluctuation power probability distribution, and automatically generate the net load fluctuation power based on the confidence interval. The method of the present invention is designed for the power fluctuation problem of the power system, and can generate expected disturbances that meet the actual scenario, providing important support for power grid scheduling and safety and stability analysis. It can be widely used in power system scheduling, new energy power generation forecasting and load fluctuation analysis and other fields. In the context of the gradual increase in the proportion of new energy, it can effectively deal with the uncertainty problems brought about by load fluctuations, and help the efficient use of clean energy and the stable operation of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] A more complete understanding of exemplary embodiments of the present invention may be obtained by referring to the following drawings:

[0071] Figure 1 Flowchart of a method 100 for adaptively generating power fluctuations of net load power in a power system according to an embodiment of the present invention;

[0072] Figure 2 A general flow chart of an online adaptive generation process of power fluctuations in a power system according to an embodiment of the present invention;

[0073] Figure 3(a), (b), and (c) thereof are respectively SSE curve graphs of wind power fluctuations, photovoltaic power fluctuations, and load fluctuations according to an embodiment of the present invention;

[0074] Figure 4 (a), (b), and (c) thereof are respectively clustering effect diagrams of wind power fluctuations, photovoltaic power fluctuations, and load fluctuations according to an embodiment of the present invention;

[0075] Figure 5 (a), (b), and (c) thereof are respectively sub - group probability distribution graphs of wind power fluctuations, photovoltaic power fluctuations, and load fluctuations according to an embodiment of the present invention;

[0076] Figure 6 is a power fluctuation disturbance power graph generated under different confidence levels throughout the day according to an embodiment of the present invention;

[0077] Figure 7 is a schematic structural diagram of an adaptive online generation system 700 for power system net load power fluctuations according to an embodiment of the present invention. Detailed Embodiments

[0078] Now, exemplary embodiments of the present invention will be introduced with reference to the accompanying drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. These embodiments are provided to disclose the present invention in detail and completely, and to fully convey the scope of the present invention to those skilled in the art. The terms in the exemplary embodiments shown in the drawings are not limitations to the present invention. In the drawings, the same unit / element uses the same reference numeral.

[0079] Unless otherwise specified, the terms (including scientific and technical terms) used herein have the ordinary meaning understood by those skilled in the art. Additionally, it can be understood that terms defined in a commonly used dictionary should be understood as having a meaning consistent with the context of their related fields, and should not be understood as idealized or overly formal meanings.

[0080] Figure 1 is a flowchart of an adaptive online generation method 100 for power system net load power fluctuations according to an embodiment of the present invention. As Figure 1As shown, the adaptive online generation method for power system net load power fluctuation provided by the embodiment of the present invention is designed for the power fluctuation problem of the power system, can generate a pre - disturbance that conforms to the actual scenario, provides important support for power grid dispatching and security and stability analysis, and can be widely applied to fields such as power system dispatching, new energy power generation prediction, and load fluctuation analysis. Under the background of the gradually increasing proportion of new energy, it can effectively cope with the uncertainty problems brought by load fluctuations, and contribute to the efficient utilization of clean energy and the stable operation of the power grid. The adaptive online generation method 100 for power system net load power fluctuation provided by the embodiment of the present invention starts from step 101. At step 101, based on the historical operation data of wind power generation power, photovoltaic power generation power, and load power with time stamps in the power grid and the weather data of the location of the power grid, feature extraction is performed to obtain feature data.

[0081] Preferably, the feature data includes three categories: power fluctuation features, meteorological features, and time features.

[0082] Among them, the power fluctuation features include: wind power active power, photovoltaic active power, load active power, wind power fluctuation power, photovoltaic fluctuation power, and load fluctuation power; the meteorological features include: temperature, wind speed, and irradiance; the time features include: month, hour, day of the week, whether it is a weekend, and whether it is a holiday.

[0083] In the present invention, multi - source data mainly includes the historical operation data of the power system and the weather data of the location of the power system. Specifically, it includes:

[0084] (1) Historical operation active power data of the power system

[0085] In the present invention, a large amount of historical data needs to be stored and analyzed. Based on the D5000 dispatching automation platform, from the.QS file in the real - time database, the power values of wind power, photovoltaic power, and load of the target power grid in real - time are obtained; from the historical database, the historical data of wind power, photovoltaic power, and load with time stamps of the target power grid are obtained, as shown in Table 1.

[0086] Table 1 Historical data table of wind power, photovoltaic power, and load with time stamps

[0087] Date_Time Value_Load Value_Wind Value_Solar 2022 / 1 / 1 14:00 62503.25 8919.25 3220.23 2022 / 1 / 1 14:05 62782.06 9079.34 3130.73 2022 / 1 / 1 14:10 62874.40 9434.02 3051.29 2022 / 1 / 1 14:15 63062.20 9719.09 2990.22 2022 / 1 / 1 14:20 63349.55 10013.81 2864.36 2022 / 1 / 1 14:25 63156.38 10402.01 2672.39 2022 / 1 / 1 14:30 63149.80 10592.15 2651.76 2022 / 1 / 1 14:35 63061.08 10794.23 2511.47 2022 / 1 / 1 14:40 63051.48 11068.21 2345.36 2022 / 1 / 1 14:45 63125.46 11148.15 2208.71 2022 / 1 / 1 14:50 63322.14 11240.25 2079.52 2022 / 1 / 1 14:55 63172.79 11366.27 1909.18 2022 / 1 / 1 15:00 63530.52 11510.89 1773.62

[0088] (2) Weather data of the location of the power grid

[0089] In the present invention, according to the geographical location of the target power grid, several locations (longitude, latitude) that can represent the weather conditions of the area where the power grid is located are selected, and historical and real - time meteorological information, including temperature, wind speed, and radiation intensity, is obtained from channels such as the National Meteorological Information Center. The weather data is shown in Table 2.

[0090] Table 2 Meteorological data table with time stamps

[0091]

[0092]

[0093] In the present invention, based on the historical wind power, photovoltaic power, load power data and meteorological data of the power grid, the extracted feature data includes: power fluctuation features, meteorological features and time features, a total of three types of features.

[0094] 1) Extraction of power and fluctuation features

[0095] In the present invention, the extracted power and fluctuation features are shown in Table 3.

[0096] Table 3 Power and fluctuation feature table

[0097]

[0098] 2) Extraction of meteorological features

[0099] In the present invention, according to the area where the power grid is located (multiple locations can be sampled), the local temperature, wind speed, and irradiance features are extracted, as shown in Table 4.

[0100] Table 4 Meteorological feature table

[0101] Serial number feature symbol meaning 1 temperature Temperature(t) The temperature or average temperature of the area where the power grid is located at time t 2 wind speed Windspeed(t) Wind speed or average wind speed in the area where the power grid is located at time t 3 irradiance Radiation(t) Irradiance or average irradiance in the area where the power grid is located at time t

[0102] 3) Time feature extraction

[0103] For time features, features such as month, hour, day of the week, whether it is a weekend, and whether it is a holiday can be extracted, as shown in Table 5.

[0104] Table 5 Time feature table

[0105]

[0106]

[0107] In step 102, the feature screening is respectively performed with the wind power fluctuation, photovoltaic power fluctuation, and load power fluctuation as the target variables, and the target features that have a great influence on the wind power fluctuation, photovoltaic power fluctuation, and load power fluctuation are screened out.

[0108] Preferably, the method further includes:

[0109] Before the feature screening is respectively performed with the wind power fluctuation, photovoltaic power fluctuation, and load power fluctuation as the target variables, the feature data is normalized.

[0110] Preferably, the feature screening is performed with the wind power fluctuation power, photovoltaic power fluctuation power, and load power fluctuation power as the target variables respectively, and the target features that have a great influence on the wind power fluctuation power, photovoltaic power fluctuation power, and load power fluctuation power are screened out, including:

[0111] Based on the tree model, the features that have a great influence on the wind power fluctuation power are screened out as the target features;

[0112] Based on PCA principal component analysis, the features that have a great influence on the photovoltaic power fluctuation power are screened out as the target features;

[0113] Based on the tree model recursive feature elimination, the features that have a great influence on the load power fluctuation power are screened out as the target features.

[0114] In the present invention, the feature data is standardized and then feature selection is performed.

[0115] In the present invention, the data standardization process includes:

[0116] ① Missing value processing: The missing values are filled by the linear interpolation method;

[0117] ② Time feature encoding: The periodic time features are encoded by sine and cosine;

[0118] ③ The feature data is Z-score standardized.

[0119] In the present invention, in order to accurately grasp the characteristics of the power grid power fluctuation disturbance, the feature correlation analysis is carried out with the wind power fluctuation power, photovoltaic power fluctuation power, and load power fluctuation power as the target variables respectively for feature selection. Among them, the wind power fluctuation is mainly closely related to features such as wind speed and irradiance. In order to effectively capture these non-linear relationships, the feature analysis is carried out based on the tree model and the features that have the greatest influence on the wind power fluctuation are screened out. The photovoltaic power fluctuation mainly has a strong collinearity with time features such as irradiance and hour sine and cosine encoding. By PCA dimensionality reduction to extract the principal components, the comprehensive information of multiple features can be captured. The load power fluctuation is mainly closely related to features such as load power and time features (such as sine and cosine encoding of the day of the week). Based on the recursive feature elimination (RFE), the feature subset is gradually optimized, and the features that have the greatest influence on the load power fluctuation are screened out as the target features.

[0120] In step 103, based on the target features corresponding to different target variables, the optimal clustering numbers corresponding to different target variables are determined, and clustering is performed based on the optimal clustering numbers to respectively determine the clustering groups corresponding to each target variable.

[0121] Preferably, the determination of the optimal clustering number corresponding to the same target variable based on the target features corresponding to different target variables includes:

[0122] Draw the SSE curves for the target features corresponding to the target variables of wind power fluctuation power, photovoltaic fluctuation power, and load fluctuation power respectively, and use the elbow method for calculation to determine the optimal number of clusters corresponding to wind power fluctuation power, photovoltaic fluctuation power, and load fluctuation power respectively.

[0123] Combined with Figure 2 As shown, in the present invention, clustering analysis of fluctuation power is performed based on historical data to determine clustering groups with wind power fluctuation power, photovoltaic fluctuation power, and load fluctuation power as target variables respectively, and the statistical features of the target variables within each clustering group are statistically analyzed.

[0124] Specifically, it includes:

[0125] (1) Determine the optimal number of clusters

[0126] The elbow method is a commonly used method for determining the optimal number of clusters. Its core idea is to find the "elbow" point of the curve by drawing the sum of squared errors (SSE) curves corresponding to different numbers of clusters k, that is, the position where the SSE decreases significantly, as the optimal number of clusters.

[0127] For a given number of clusters k, the calculation formula of SSE is:

[0128]

[0129] Among them, C i represents all sample points in the i-th cluster; μ i represents the center point of the i-th cluster; x represents a certain sample point belonging to the i-th cluster.

[0130] As k increases, SSE will gradually decrease. When k reaches a certain value, the decrease rate of SSE will significantly slow down, forming an "elbow". Plot the change of SSE with k as a curve and observe the "elbow" position of the curve. Select the k value corresponding to the "elbow" point as the optimal number of clusters.

[0131] (2) Conduct clustering analysis of fluctuation power

[0132] K-Means++ is an improved K-Means clustering initialization method, which solves the problem that random initialization of cluster centers in the traditional K-Means algorithm may lead to slow convergence speed or unstable results.

[0133] The clustering steps based on the K-Means++ algorithm are as follows:

[0134] 1) Initialize the first center point: Randomly select a point from the dataset as the first cluster center.

[0135] 2) Select subsequent central points: For each data sample, calculate the distance D(x) from it to the nearest selected central point, and select subsequent central points with probability P(x) ∝ D(x).

[0136] 3) Assign samples: Assign each sample to the group to which the nearest central point belongs.

[0137] 4) Update central points: Calculate the mean of each group and update it as the new central point.

[0138] 5) Iterative optimization: Repeat steps 2 and 3 until the central points no longer change significantly or the maximum number of iterations is reached.

[0139] 6) Save the clustering results: Save the normalizer, cluster centers, and cluster labels, and add the cluster labels to the original dataset.

[0140] (3) Analyze the clustering results

[0141] For target variables such as wind power fluctuation, photovoltaic power fluctuation, and load power fluctuation respectively, calculate the statistical characteristics of the target variable corresponding to each cluster group result within the group, including: mean, standard deviation, minimum value, maximum value, and store the results.

[0142] In step 104, based on the real-time operation data and weather data of the power grid, feature extraction is performed to obtain real-time feature samples, calculate the Euclidean distance between the real-time feature samples and the cluster centers of each cluster group, and determine the target cluster group corresponding to each target variable based on the nearest Euclidean distance.

[0143] Preferably, the method further includes:[[]]

[0144] Perform normalization processing on the real-time feature samples.

[0145] In the present invention, in order to generate power fluctuations, it is also necessary to obtain the real-time data of power grid meteorology-time-power and meteorological features, date features, and integrate them to form real-time samples; then load the normalizer saved during the power fluctuation clustering analysis process, and perform preprocessing and normalization similar to historical data on the real-time samples to obtain real-time feature samples.

[0146] In the present invention, for real-time feature samples, the identification of the group to which they belong is performed to determine the target clustering group. Specifically, for the formed standardized real-time data samples, for the clustering results of wind power fluctuation power, photovoltaic power fluctuation power, and the clustering analysis results of load fluctuation power, the Euclidean distances between the real-time feature samples and the clustering centers of each clustering group are calculated respectively; finally, for the clustering result corresponding to any target variable, the clustering group with the smallest Euclidean distance is selected as the target clustering group corresponding to any target variable.

[0147] Among them, the Euclidean distance formula is: d i = ||x - μ i ||.

[0148] In step 105, the statistical characteristics of the target variables corresponding to each target clustering group are calculated respectively, and the probability distribution of the net load fluctuation power is calculated based on the statistical characteristics.

[0149] Preferably, calculating the probability distribution of the net load fluctuation power based on the statistical characteristics includes:

[0150] Based on the means and variances of the target variables corresponding to the target clustering groups of wind power fluctuation power, photovoltaic power fluctuation power, and load fluctuation power, a covariance matrix and a joint probability distribution are constructed respectively; among them, the covariance matrix is:

[0151]

[0152] The joint probability distribution is:

[0153]

[0154] Determining the probability distribution of the net load fluctuation power based on the variance matrix and the joint probability distribution includes:

[0155]

[0156] Among them, ΔP_NetLoad is the net load fluctuation power; and are respectively the mean and standard deviation of the wind power fluctuation power of the target clustering group corresponding to the wind power fluctuation power; and are respectively the mean and standard deviation of the photovoltaic power fluctuation power of the target clustering group corresponding to the photovoltaic power fluctuation power; and are the mean and standard deviation of the photovoltaic power fluctuation of the target cluster group corresponding to the load fluctuation power; CM is the covariance matrix; ΔP_Wind, ΔP_Solar and ΔP_Load are the wind charge fluctuation active power, photovoltaic fluctuation active power and load fluctuation power respectively; cov() represents the correlation coefficient calculation function; μ net and σ net are the mean and standard deviation of net load fluctuation power, respectively.

[0157] In the present invention, after real-time data samples are matched to their respective power fluctuation groups, it is necessary to further analyze the correlation between wind power, photovoltaic power, and load fluctuation power, and determine the net load fluctuation power probability distribution based on this. The specific steps are as follows:

[0158] 1) Extract the statistical summary of the matching group. The index numbers of the wind power fluctuation, photovoltaic fluctuation and load fluctuation group matched by the real-time data are i and w 、i s 、i l , extract the mean and standard deviation of each group. Specifically:

[0159] ① Wind power fluctuation i w The statistical summary of the group is: Mean Standard deviation

[0160] ② Photovoltaic fluctuation i s The statistical summary of the group is: Mean Standard deviation

[0161] ③ Load fluctuation i l The statistical summary of the group is: Mean Standard deviation

[0162] 2) Construct the covariance matrix to calculate the wind power fluctuation i w Group, photovoltaic fluctuations i s Group and load fluctuations l The covariance matrix of the fluctuating power between groups. The formula for calculating the covariance matrix is as follows:

[0163]

[0164] 3) Joint Probability Distribution Modeling Assuming that wind power fluctuations, photovoltaic fluctuations, and load fluctuations obey multivariate normal distributions, the probability distribution of the fluctuating power is expressed as:

[0165]

[0166] 4) Calculation of Net Load Fluctuation Power Probability Distribution In the present invention, net load fluctuation is defined as:

[0167] ΔP_NetLoad = ΔP_Load - (ΔP_Wind + ΔP_Solar)

[0168] The mean and variance of the net load fluctuation power can be derived from the joint distribution:

[0169] The mean is:

[0170] The variance:

[0171]

[0172] Therefore, the probability distribution of the net load power fluctuation can be expressed as:

[0173]

[0174] where ΔP_NetLoad is the net load fluctuation power; and are the mean and standard deviation of the wind power fluctuation corresponding to the target clustering group of the wind power fluctuation respectively; and are the mean and standard deviation of the photovoltaic power fluctuation corresponding to the target clustering group of the photovoltaic power fluctuation respectively; and are the mean and standard deviation of the photovoltaic power fluctuation corresponding to the target clustering group of the load fluctuation power respectively; CM is the covariance matrix; ΔP_Wind, ΔP_Solar and ΔP_Load are the active power fluctuations of wind power, photovoltaic power and load respectively; cov() represents the correlation coefficient calculation function; μ net and σ net are the mean and standard deviation of the net load fluctuation power respectively.

[0175] In step 106, based on the probability distribution of the net load fluctuation power, determine the confidence interval of the net load fluctuation power, and automatically generate the net load fluctuation power based on the confidence interval.

[0176] Preferably, determining the confidence interval of the net load fluctuation power based on the probability distribution of the net load fluctuation power includes:

[0177] Determine a preset confidence level, and convert the preset confidence level to a Z value according to the standard normal distribution table; <�

[0178] Determine that the confidence interval of the net load fluctuation power under the preset confidence level is: [μ net - Z·σ net , μ net + Z·σ net ; where μ netand σ net are the mean and standard deviation of the net load fluctuation power respectively; Z represents the Z value corresponding to the preset confidence level.

[0179] The confidence interval is a range for estimating the population parameter based on sample data. For the case of normal distribution, the upper and lower bounds of the confidence interval (denoted as L and U respectively) can be calculated by the following formula:

[0180] L = μ - Z·σ

[0181] U = μ + Z·σ.

[0182] In the present invention, first, according to the probability distribution of the net load power fluctuation, the confidence interval ranges under different confidence levels can be generated. For example, the confidence level Confidence_Level = 99.0% is set.

[0183] Then, according to the standard normal distribution table, the confidence level is converted into the corresponding Z value:

[0184]

[0185] Finally, the confidence interval is determined as: [μ net [[ID=--Z·σ net , μ net +Z·σ net , and based on the said confidence interval, the automatic generation of the net load fluctuation power can be carried out; wherein, μ net and σ net are the mean and standard deviation of the net load fluctuation power respectively; Z represents the Z value corresponding to the preset confidence level.

[0186] In summary, the present invention integrates the meteorological characteristics and time characteristics based on the wind power, photovoltaic power and load fluctuation power, and forms a historical operation data sample set of the power system meteorology - time - power; takes the wind power, photovoltaic power and load fluctuation power as the research objectives respectively, and realizes the separate clustering of the fluctuation power; the real - time sample determines the index number of the affiliated clustering group by matching the closest clustering cluster; in order to further describe the correlation between the wind power, photovoltaic power and load fluctuation, the covariance matrix and joint probability distribution of the clustering group to which the current sample belongs are calculated, and then the probability distribution of the net load power fluctuation is calculated, and the net load fluctuation power under different confidence levels is generated through the net load probability distribution. The generated net load fluctuation interval can support the power grid to carry out online safety early warning, online checking of the starting mode and optimization of the reserve capacity, etc. It provides a basis for disturbance setting for realizing the power generation - consumption balance analysis and dispatching control at the minute - to - hour level.

[0187] The effects of the present invention are as follows:

[0188] (1) Innovation in the disturbance generation method

[0189] A method for generating expected disturbances suitable for time scales from minutes to hours is proposed, which provides a high-quality expected disturbance set basis for long-time scale safety analysis.

[0190] (2) Multi-feature fusion improves accuracy

[0191] The introduction of meteorological features (such as temperature, wind speed, radiation) and time features (such as hour, week, month) significantly enhances the accuracy of online generation of disturbance morphology.

[0192] (3) Multi-dimensional decision support

[0193] A fluctuation range with different confidence levels (such as 99%, 99.9%, etc.) is generated for each group, providing a quantitative basis for risk assessment and scientific decision-making of the power system.

[0194] (4) Identification of key factors

[0195] Through in-depth analysis of group characteristics, the key driving factors affecting power system fluctuations (such as weather conditions, time periods, etc.) can be effectively identified to provide guidance for system optimization.

[0196] (5) Wide application value

[0197] It can be widely used in power system scheduling, renewable energy generation forecasting, and load fluctuation analysis. As the proportion of renewable energy gradually increases, it can effectively address the uncertainty caused by load fluctuations, helping to achieve efficient utilization of clean energy and stable grid operation.

[0198] The following examples illustrate the embodiments of the present invention.

[0199] In an embodiment of the present invention, the historical operating data of a certain actual regional power grid in 2022 is used as an example of the above-mentioned scheme to illustrate the effectiveness of the present invention.

[0200] Specifically, the process of adaptive online generation of power system net load power fluctuations includes:

[0201] (1) Multi-source data fusion and feature engineering

[0202] The historical data of wind power, photovoltaic power and load from 2022-01-01 00:05:00 to 2022-12-30 23:55:00 are used to integrate meteorological data and time characteristics, extract data features and perform standardization.

[0203] After processing, the historical dataset integrating weather-time-power correlation characteristics includes the following characteristic variables:

[0204] 1) Wind power, photovoltaic power, load power and its fluctuating power, a total of 6 power characteristic variables;

[0205] 2) Set 1 meteorological data (including temperature, wind speed, and irradiance) sampling point in each of the 4 provinces within the scope of the regional power grid, for a total of 12 meteorological characteristic variables.

[0206] 3) In terms of time characteristics, in addition to the characteristics of "whether it is a weekend" and "whether it is a holiday", perform sine and cosine encoding on the periodic characteristics (month, hour, day of the week), and a total of 8 time characteristic variables are obtained.

[0207] Taking the wind power fluctuation power, photovoltaic power fluctuation power, and load fluctuation power as the target variables respectively, conduct feature correlation analysis. In order to reduce the dimension of the dataset, average the meteorological data, and use the average temperature, average wind speed, and average irradiance as features.

[0208] The feature selection results are as follows. Based on the tree model, 5 strongly correlated features of wind power fluctuation power are selected, as shown in Table 6; based on PCA principal component analysis, 6 strongly correlated features of photovoltaic power fluctuation power are selected, as shown in Table 7; based on the recursive feature elimination of the tree model, 8 strongly correlated features of load fluctuation power are selected, as shown in Table 8.

[0209] Table 6 Feature Selection Table for Wind Power Fluctuation Power

[0210] Serial number feature symbol Remark Wind power fluctuation ΔP_Wind Target variable 1 Regional average temperature Avg_Temperature 2 Wind power Value_Wind 3 Regional average wind speed Avg_Windspeed 4 Regional average irradiance Avg_Radiation 5 Hour sine code Hour_sin

[0211] Table 7 Feature Selection Table for Photovoltaic Power Fluctuation Power

[0212] Serial number feature symbol Remark Photovoltaic fluctuating power ΔP_Solar Target variable 1 Regional average irradiance Avg_Radiation 2 Photovoltaic power Value_Solar 3 Hour cosine encoding Hour_cos 4 Regional average temperature Avg_Temperature 5 Month cosine code Month_cos 6 Regional average wind speed Avg_Windspeed

[0213] Table 8 Feature Selection Table for Load Fluctuation Power

[0214]

[0215]

[0216] (2) Determine the optimal number of clusters and K-Means++ clustering

[0217] 1) Determine the optimal number of clusters

[0218] After selecting the main features, taking the wind power fluctuation power, photovoltaic power fluctuation power, and load fluctuation power as the target variables respectively, draw the SSE curve, as shown in (a), (b), and (c) of Figure 3 According to the elbow method, the optimal number of clusters for wind power fluctuation power, photovoltaic power fluctuation power, and load fluctuation power is 4.

[0219] 2) Conduct clustering analysis of power fluctuations

[0220] Cluster using the K-Means++ algorithm. After completion, save the normalizer, cluster centers, and cluster labels, and add the cluster labels to the original dataset. The clustering results are shown in (a), (b), and (c) of Figure 4 respectively.

[0221] 3) Analyze the clustering results

[0222] For target variables such as wind power fluctuations, photovoltaic power fluctuations, and load power fluctuations, calculate the statistical characteristics (mean, standard deviation, minimum, maximum) of the target variables within each cluster group for each variable, and store the results. The statistical characteristic results are shown in (a), (b), and (c) of Figure 5 respectively, and the specific data are shown in Table 9, Table 10, and Table 11 respectively.

[0223] Table 9 Statistical Characteristics of Each Group of Wind Power Fluctuations

[0224] Wind_Cluster Mean Std Min Max 0 -34.43 322.25 -2036.12 2072.62 1 111.81 238.74 -1281.16 1647.47 2 64.60 261.98 -1266.12 1356.68 3 -107.41 221.68 -1502.06 879.85

[0225] Table 10 Statistical Characteristics of Each Group of Photovoltaic Power Fluctuations

[0226] Solar_Cluster Mean Std Min Max 0 589.14 333.66 -343.77 2341.67 1 -48.77 234.30 -1993.70 714.02 2 -229.89 363.41 -2163.40 772.77 3 -2.08 187.68 -1307.45 737.21

[0227] Table 11 Statistical Characteristics of Each Group of Load Power Fluctuations

[0228] Load_Cluster Mean Std Min Max 0 220.35 464.89 -957.42 2169.63 1 -303.25 407.71 -1630.36 1246.19 2 10.51 516.22 -1427.17 2411.61 3 55.65 358.76 -1597.11 1533.76

[0229] 4) Calculate the covariance matrices of wind power fluctuations, photovoltaic power fluctuations, and load power fluctuations

[0230] At this time, calculate the covariance matrices of wind, photovoltaic, and load power fluctuations. Each type of fluctuation variable is divided into 4 groups, so there are at most 4x4x4 = 64 combinations. Each combination corresponds to a covariance matrix. In the data structure, there are three label columns: Wind_Cluster, Solar_Cluster, and Load_Cluster, each with an index from 0 to 3. The three fluctuation power data columns, ΔP_wind, ΔP_solar, and ΔP_load, are the corresponding fluctuation power data. The covariance matrix needs to be calculated for each group combination of these three variables. For example, when Wind_Cluster = 0, Solar_Cluster = 0, and Load_Cluster = 0, the fluctuation data of these three variables are extracted and their covariance matrix is calculated, as shown in Table 11.

[0231] Table 12 Covariance Matrix Table for Wind_Cluster = 0, Solar_Cluster = 0, Load_Cluster = 0

[0232] ΔP_wind ΔP_solar ΔP_load ΔP_wind 148621.95 -25455.96 30964.97 ΔP_solar -25455.96 42744.21 -8488.40 ΔP_load 30964.97 -8488.40 101980.23

[0233] (3) Online Generation of Anticipated Disturbances for Power Fluctuations

[0234] 1) Formation of Real-Time Data of Grid Meteorology - Time - Power

[0235] Taking the data from 00:05:00 on December 31, 2022 to 23:55:00 on December 31, 2022 as real-time operation data, during the model establishment process, the data on December 31, 2022 was not added to the training to avoid future sample leakage into the model data. Load the real-time grid data, obtain meteorological features and date features, and integrate them to form real-time samples; and load the normalizer saved during the clustering analysis of fluctuating power, and perform preprocessing and normalization similar to historical data on the real-time samples.

[0236] 2) Identification of the Cluster to Which the Current Real-Time Sample Point Belongs

[0237] Taking the data sample at 00:05:00 on December 31, 2022 as an example, analyze the clusters to which it belongs in the clustering results of wind power fluctuations, photovoltaic power fluctuations, and load power fluctuations in turn.

[0238] The results are as follows:

[0239] Processing Wind Fluctuations:

[0240] Closest Cluster Index:0

[0241] Processing Solar Fluctuations:

[0242] Closest Cluster Index:1

[0243] Processing Load Fluctuations:

[0244] Closest Cluster Index:2.

[0245] (4) Calculation of the Joint Probability Distribution of Wind Power Fluctuations, Photovoltaic Power Fluctuations, and Load Fluctuations

[0246] After the real-time data samples are matched to their respective power fluctuation groups, it is necessary to further analyze the correlation between wind power, photovoltaic power, and load fluctuation power. The specific steps are as follows:

[0247] 1) Extract the Statistical Summary of the Matched Group

[0248] The group index numbers of the wind power fluctuations, photovoltaic fluctuations, and load fluctuations matched with real-time data are 0, 1, and 2 respectively. The mean and standard deviation of each group are extracted. Specifically:

[0249] ① The statistical summary of the wind power fluctuation group 0 is: mean μ wind,0 =-34.43; standard deviation σ wind,0 =322.25;

[0250] ② The statistical summary of the photovoltaic fluctuation group 1 is: mean μ solar,1 =-48.77; standard deviation σ solar,1 =234.30;

[0251] ③ The statistical summary of the load fluctuation group 2 is: mean μ load,2 =10.51; standard deviation σ load,2 =516.22.

[0252] 2) Read the covariance matrix

[0253] Calculate the covariance matrix of the fluctuation power between the wind power fluctuation group 0, the photovoltaic fluctuation group 1, and the load fluctuation group 2. The covariance matrix is as follows:

[0254]

[0255] 3) Joint probability distribution modeling

[0256] Assume that the wind power fluctuations, photovoltaic fluctuations, and load fluctuations follow a multivariate normal distribution. Then the joint probability distribution followed by the data sample fluctuation power at 00:05:00 on December 31, 2022 is expressed as:

[0257]

[0258] (5) Calculation of the probability distribution of the net load fluctuation power

[0259] The net load fluctuation is defined as

[0260] ΔP_NetLoad = ΔP_Load - (ΔP_Wind + ΔP_Solar)

[0261] The mean and variance of the net load fluctuation power can be derived from the joint distribution. For the real-time data sample at 00:05:00 on December 31, 2022:

[0262] Mean:

[0263] Variance:

[0264]

[0265] Therefore, the probability distribution of the net load power fluctuation at the moment of 2022-12-31 00:05:00 can be expressed as:

[0266] ΔP_NetLoad~N(93.71,730.51 2 )

[0267] (6) Calculation of the net load power fluctuation interval under multiple confidence levels of real-time sample points

[0268] 1) According to the probability distribution of the net load power fluctuation, the confidence interval ranges under different confidence levels can be generated. For example, set the confidence level Confidence_Level = 99.0%.

[0269] 2) According to the standard normal distribution table, convert the confidence level to the corresponding Z value

[0270]

[0271] Then, the confidence interval calculation formula is expressed as:

[0272] [μ net -Z·σ net ,μ net +Z·σ net ,

[0273] where μ net and σ net are the mean and standard deviation of the net load fluctuation power respectively; Z represents the Z value corresponding to the preset confidence level.

[0274] The confidence interval is a range for estimating the population parameter based on sample data. For the case of normal distribution, the upper and lower bounds of the confidence interval (denoted as L and U respectively) can be calculated by the following formula:

[0275]

[0276] Finally, adopting the foregoing scheme, for a large amount of historical actual operation data, the fluctuation power verification under different confidence levels is carried out. The predicted disturbance of the net load power fluctuation generated by the scheme proposed by the present invention can cover the fluctuation disturbance power actually occurring in the power grid. The power fluctuation disturbance value determined by the present invention does not show a serious deviation from the actual power curve, and can effectively avoid the over-allocation of reserves such as adjustable power sources, thereby improving the economic efficiency of the power grid operation. The generation of the power fluctuation disturbance power under different confidence levels throughout the day on December 31, 2022 is as Figure 6 shown

[0277] Figure 7 is a schematic structural diagram of the power system net load power fluctuation adaptive online generation system 700 according to an embodiment of the present invention. As Figure 7 As shown in Figure 7 , the adaptive online generation system 700 for power system net load power fluctuation provided by the embodiment of the present invention includes: a feature data acquisition unit 701, a feature screening unit 702, a clustering unit 703, a target clustering group determination unit 704, a probability distribution determination unit 705, and a confidence interval determination unit 706.

[0278] Preferably, the feature data acquisition unit 701 is configured to perform feature extraction based on the historical operation data of wind power generation power, photovoltaic power generation power, and load power with time stamps in the power grid and the weather data of the location of the power grid to obtain feature data.

[0279] Preferably, the feature data includes three categories: power fluctuation features, meteorological features, and time features;

[0280] Among them, the power fluctuation features include: wind power active power, photovoltaic active power, load active power, wind power fluctuation power, photovoltaic fluctuation power, and load fluctuation power; the meteorological features include: temperature, wind speed, and irradiance; the time features include: month, hour, day of the week, whether it is a weekend, and whether it is a holiday.

[0281] Preferably, the feature screening unit 702 is configured to perform feature screening with wind power fluctuation power, photovoltaic fluctuation power, and load fluctuation power as target variables respectively, and screen out target features that have a great influence on wind power fluctuation power, photovoltaic fluctuation power, and load fluctuation power.

[0282] Preferably, the system further includes:

[0283] A normalization processing unit, configured to perform normalization processing on the feature data before performing feature screening with wind power fluctuation power, photovoltaic fluctuation power, and load fluctuation power as target variables respectively.

[0284] Preferably, the feature screening unit 703 performs feature screening with wind power fluctuation power, photovoltaic fluctuation power, and load fluctuation power as target variables respectively, and screens out target features that have a great influence on power fluctuation power, photovoltaic fluctuation power, and load fluctuation power, including:

[0285] Based on a tree model, screen out features that have a great influence on wind power fluctuation power as target features;

[0286] Based on PCA principal component analysis, screen out features that have a great influence on photovoltaic fluctuation power as target features;

[0287] Based on tree model recursive feature elimination, screen out features that have a great influence on load fluctuation power as target features.

[0288] Preferably, the clustering unit 703 is configured to determine the optimal number of clusters corresponding to different target variables based on the target features corresponding to different target variables, and perform clustering based on the optimal number of clusters to respectively determine the clustering groups corresponding to each target variable.

[0289] Preferably, the clustering unit 703 determines the optimal number of clusters corresponding to the same target variable based on the target features corresponding to different target variables, including:

[0290] Based on the wind power fluctuation, photovoltaic power fluctuation, and load power fluctuation as the target features corresponding to the target variables, draw SSE curves respectively, and use the elbow method for calculation to respectively determine the optimal number of clusters corresponding to the wind power fluctuation, photovoltaic power fluctuation, and load power fluctuation.

[0291] Preferably, the target clustering group determination unit 704 is configured to extract features based on the real-time operation data and weather data of the power grid to obtain real-time feature samples, calculate the Euclidean distance between the real-time feature samples and the cluster centers of each clustering group, and determine the target clustering group corresponding to each target variable based on the nearest Euclidean distance.

[0292] Preferably, the system further includes:

[0293] A normalization processing unit, configured to perform normalization processing on the real-time feature samples.

[0294] Preferably, the probability distribution determination unit 705 is configured to calculate the statistical features of the target variables corresponding to each target clustering group respectively, and calculate the net load power fluctuation probability distribution based on the statistical features.

[0295] Preferably, the probability distribution determination unit 705 calculates the net load power fluctuation probability distribution based on the statistical features, including:

[0296] Based on the mean and variance of the target variables corresponding to the target clustering groups corresponding to the wind power fluctuation, photovoltaic power fluctuation, and load power fluctuation, construct a covariance matrix and a joint probability distribution respectively; where the covariance matrix is:

[0297]

[0298] The joint probability distribution is:

[0299]

[0300] Determining the probability distribution of the net load power fluctuation based on the variance matrix and the joint probability distribution, including:

[0301]

[0302] Among them, ΔP_NetLoad is the net load fluctuation power; and are the mean and standard deviation of the wind power fluctuation power of the target cluster group corresponding to the wind power fluctuation power; and are the mean and standard deviation of the photovoltaic fluctuation power of the target cluster group corresponding to the photovoltaic fluctuation power; and are the mean and standard deviation of the photovoltaic power fluctuation of the target cluster group corresponding to the load fluctuation power; CM is the covariance matrix; ΔP_Wind, ΔP_Solar and ΔP_Load are the wind charge fluctuation active power, photovoltaic fluctuation active power and load fluctuation power respectively; cov() represents the correlation coefficient calculation function; μ net and σ net are the mean and standard deviation of net load fluctuation power, respectively.

[0303] Preferably, the confidence interval determining unit 706 is configured to determine a confidence interval of the net load fluctuation power based on the net load fluctuation power probability distribution, and automatically generate the net load fluctuation power based on the confidence interval.

[0304] Preferably, the probability distribution determining unit 706 determines the confidence interval of the net load fluctuation power based on the net load fluctuation power probability distribution, including:

[0305] Determining a preset reliability, and converting the preset reliability into a Z value according to a standard normal distribution table;

[0306] The confidence interval of the net load fluctuation power under the preset confidence level is determined as: [μ net -Z·σ net ,μ net +Z·σ net ]; where μ net and σ net are the mean and standard deviation of the net load fluctuation power respectively; Z represents the Z value corresponding to the preset confidence level.

[0307] The power system net load power fluctuation adaptive online generation system 700 of the embodiment of the present invention corresponds to the power system net load power fluctuation adaptive online generation method 100 of another embodiment of the present invention, and will not be repeated here.

[0308] According to another aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored. When the program is executed by a processor, any step of a method for adaptive online generation of net load power fluctuations in a power system is implemented.

[0309] On the other hand of the present invention, the present invention provides an electronic device, including:

[0310] The above-mentioned computer-readable storage medium; and one or more processors for executing the program in the computer-readable storage medium.

[0311] The present invention has been described by referring to a few embodiments. However, as is well known to those skilled in the art, other embodiments equivalent to those disclosed above of the present invention equally fall within the scope of the present invention.

[0312] Generally, all terms used in the present invention are interpreted according to their ordinary meanings in the technical field, unless otherwise clearly defined therein. All references to "a / the [device, component, etc.]" are to be interpreted openly as at least one instance of the device, component, etc., unless otherwise clearly stated. The steps of any method disclosed herein need not be run in the exact order disclosed, unless clearly stated.

[0313] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) containing computer-usable program code.

[0314] The present application is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for realizing the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0315] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device that realizes the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0316] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, so that a series of operation steps are executed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing instructions for implementing the steps specified in one process or multiple processes and / or blocks Figure 1 one process or multiple processes and / or blocks Figure 1 in one block or multiple blocks.

[0317] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications or equivalent replacements can still be made to the specific embodiments of the present invention. Any modification or equivalent replacement without departing from the spirit and scope of the present invention shall be covered by the protection scope of the present invention.

Claims

1. An adaptive online generation method for power system net load power fluctuations, characterized in that, The method includes: Performing feature extraction based on the historical operation data of wind power generation power, photovoltaic power generation power, and load power with time stamps in the power grid and the weather data at the location of the power grid to obtain feature data; Performing feature screening with the wind power fluctuation power, photovoltaic power fluctuation power, and load power fluctuation power as target variables respectively, and screening out the target features that have a great influence on the wind power fluctuation power, photovoltaic power fluctuation power, and load power fluctuation power; Determining the optimal number of clusters corresponding to different target variables based on the target features corresponding to different target variables, and performing clustering based on the optimal number of clusters to respectively determine the clustering groups corresponding to each target variable; Performing feature extraction based on the real-time operation data and weather data of the power grid to obtain real-time feature samples, calculating the Euclidean distance between the real-time feature samples and the cluster centers of each clustering group, and determining the target clustering group corresponding to each target variable based on the nearest Euclidean distance; Calculating the statistical features of the target variables corresponding to each target clustering group respectively, and calculating the probability distribution of the net load fluctuation power based on the statistical features; Determining the confidence interval of the net load fluctuation power based on the probability distribution of the net load fluctuation power, and automatically generating the net load fluctuation power based on the confidence interval.

2. The method according to claim 1, wherein The feature data includes three categories: power fluctuation features, meteorological features, and time features; Among them, the power fluctuation features include: wind power active power, photovoltaic active power, load active power, wind power fluctuation power, photovoltaic power fluctuation power, and load power fluctuation power; the meteorological features include: temperature, wind speed, and irradiance; the time features include: month, hour, day of the week, whether it is a weekend, and whether it is a holiday.

3. The method according to claim 1, wherein The method further includes: Before performing feature screening with the wind power fluctuation power, photovoltaic power fluctuation power, and load power fluctuation power as target variables respectively, performing standardization processing on the feature data.

4. The method according to claim 1, wherein The performing feature screening with the wind power fluctuation power, photovoltaic power fluctuation power, and load power fluctuation power as target variables respectively, and screening out the target features that have a great influence on the wind power fluctuation power, photovoltaic power fluctuation power, and load power fluctuation power includes: Screening out the features that have a great influence on the wind power fluctuation power based on a tree model as target features; Screening out the features that have a great influence on the photovoltaic power fluctuation power based on PCA principal component analysis as target features; Screening out the features that have a great influence on the load power fluctuation power based on the recursive feature elimination of a tree model as target features.

5. The method according to claim 1, characterized in that, The determining the optimal number of clusters corresponding to different target variables based on the target features corresponding to different target variables includes: Respectively plotting SSE curves based on the target features corresponding to the wind power fluctuation power, photovoltaic power fluctuation power, and load power fluctuation power as target variables, and calculating using the elbow method to respectively determine the optimal number of clusters corresponding to the wind power fluctuation power, photovoltaic power fluctuation power, and load power fluctuation power.

6. The method according to claim 1, characterized in that, The method further includes: Performing standardization processing on the real-time feature samples.

7. The method according to claim 1, wherein The calculating the probability distribution of the net load fluctuation power based on the statistical features includes: Based on the means and variances of the target variables corresponding to the wind power fluctuation, photovoltaic power fluctuation, and load power fluctuation in the corresponding target clustering groups, a covariance matrix and a joint probability distribution are constructed respectively; among them, the covariance matrix is: The joint probability distribution is: Determining the probability distribution of the net load power fluctuation based on the variance matrix and the joint probability distribution includes: ΔP_NetLoad = ΔP_Load - (ΔP_Wind + ΔP_Solar) Among them, ΔP_NetLoad is the net load fluctuation power; and are the mean and standard deviation of the wind power fluctuation of the target clustering group corresponding to the wind power fluctuation, respectively; and are the mean and standard deviation of the PV power fluctuation of the target clustering group corresponding to the PV power fluctuation, respectively; and are the mean and standard deviation of the PV power fluctuation of the target clustering group corresponding to the load fluctuation power, respectively; CM is the covariance matrix; ΔP_Wind, ΔP_Solar, and ΔP_Load are the active power fluctuations of wind power, PV power, and load fluctuation, respectively; cov() represents the correlation coefficient calculation function; μ net and σ net are the mean and standard deviation of the net load fluctuation power, respectively.

8. The method according to claim 1, wherein Determining the confidence interval of the net load power fluctuation based on the net load power fluctuation probability distribution includes: Determining a preset confidence level and converting the preset confidence level into a Z value according to the standard normal distribution table; The confidence interval for the net load fluctuation power at the preset confidence level is: [μ net -Z·σ net , μ net +Z·σ net ; where μ net and σ net are the mean and standard deviation of the net load fluctuation power respectively; Z represents the Z value corresponding to the preset confidence level.

9. An adaptive online generation system for power system net load power fluctuations, characterized in that, The system includes: A feature data acquisition unit, configured to perform feature extraction based on the historical operation data of the wind power generation power, photovoltaic power generation power, and load power with time stamps in the power grid and the weather data at the location of the power grid, and acquire feature data; A feature screening unit, configured to perform feature screening with the wind power fluctuation, photovoltaic power fluctuation, and load power fluctuation as target variables respectively, and screen out the target features that have a great influence on the wind power fluctuation, photovoltaic power fluctuation, and load power fluctuation; A clustering unit, configured to determine the optimal number of clusters corresponding to different target variables based on the target features corresponding to different target variables, and perform clustering based on the optimal number of clusters to respectively determine the clustering groups corresponding to each target variable; A target clustering group determination unit, configured to perform feature extraction based on the real-time operation data and weather data of the power grid to obtain real-time feature samples, calculate the Euclidean distances between the real-time feature samples and the cluster centers of each clustering group, and determine the target clustering groups corresponding to each target variable based on the nearest Euclidean distance; A probability distribution determination unit, configured to calculate the statistical features of the target variables corresponding to each target clustering group respectively, and calculate the net load power fluctuation probability distribution based on the statistical features; A confidence interval determination unit, configured to determine the confidence interval of the net load power fluctuation based on the net load power fluctuation probability distribution, and perform automatic generation of the net load power fluctuation based on the confidence interval.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by a processor, it implements the steps of the method described in any one of claims 1-8.

11. An electronic device, characterized in that, Including: The computer-readable storage medium described in claim 10; And One or more processors, configured to execute the program in the computer-readable storage medium.

Images