Method and system for evaluating operational reserve capacity requirements considering uncertainty of new energy sources

Abstract

This application discloses a method and system for assessing operational reserve capacity demand that takes into account the uncertainty of new energy sources. The method comprises: constructing a time series of comprehensive influencing factors of new energy forecast errors and a time series of comprehensive influencing factors of load forecast errors for the planned day based on the new energy forecast data and load forecast data for the planned day; selecting the new energy sub-database and load sub-database with the greatest similarity to determine them as the new energy sub-database and load sub-database for the planned day; constructing a probability distribution of new energy forecast errors and a probability distribution of load forecast errors for the new energy sub-database and the load sub-database, respectively, to obtain a net load forecast probability distribution; and calculating the operational reserve capacity demand based on the net load forecast probability distribution and a preset confidence level. This application considers the impact of factors such as source-load bidirectionality, meteorological conditions, and power on forecast errors, thereby improving the accuracy of reserve demand assessment.

Description

Technical Field

[0001] The present invention belongs to the field of operation scheduling, and in particular relates to an operation reserve capacity demand assessment method and system taking into account the uncertainty of new energy. Background Art

[0002] With the rapid development and widespread grid integration of renewable energy, the optimal dispatch of power systems containing a high proportion of renewable energy has become a highly sought-after research topic in the current energy landscape. As a crucial factor in ensuring the safe and stable operation of power systems, the optimization of backup resources also faces new challenges in the context of large-scale grid integration of wind power.

[0003] The purpose of reserving reserve in power systems is to mitigate power imbalances caused by uncertainties on the source and load sides of the system. Excessive reserve leads to waste and reduced economic efficiency, while too little reserve is insufficient to ensure reliable system operation. Therefore, an accurate and reasonable assessment of reserve requirements is necessary to achieve a balance between system reliability and economic efficiency. Before the integration of large-scale wind power and flexible loads, reserve requirements were assessed using deterministic methods, often employing the N-1 criterion and load percentage criteria. However, with the integration of a high proportion of wind power, the relationship between system uncertainty and reserve requirements becomes complex, making deterministic methods no longer applicable.

[0004] Prior art document 1 (CN118917574A) discloses a method and system for evaluating the energy backup demand of a power grid system taking into account time series characteristics. A dynamic time warping method is used to obtain a new energy forecast curve for a historical date that is similar in shape to the new energy forecast curve for the planned date, and the historical date is added to a historical sample date set; based on the dates in the historical sample date set, the difference between the new energy forecast curve and the new energy measured curve for each time period of the date is calculated to generate a historical sample data sequence of the new energy forecast error; the time periods in the historical sample data sequence are clustered to obtain aggregated time periods; the probability density function of the new energy forecast error for the aggregated time period is calculated; and finally, the new energy backup demand is calculated based on the set confidence level and the probability density function. Prior art document 2 (CN119448444A) discloses a grid reserve capacity adjustment method and system that considers source-load correlation, including: obtaining the predicted value of renewable energy power generation and grid load in each time period to determine the predicted value of net load in each time period; determining the net load of each typical scenario in each time period based on the renewable energy power generation and grid load in each typical scenario in each time period; when the weighted sum of the difference between the net load of each typical scenario in each time period and the predicted net load value and the probability of each typical scenario is greater than zero, using the weighted sum as the adjusted grid reserve capacity in each time period; using the larger value of the grid reserve capacity before and after adjustment in each time period as the preferred value of the grid reserve capacity in each time period; and adjusting the grid reserve capacity according to the preferred value of the grid reserve capacity in each time period.

[0005] However, prior art document 1 only considers the uncertainty of renewable energy forecasts, failing to effectively account for the impact of uncertainty on both the source and load sides. Furthermore, this method only considers the impact of forecast randomness through historical similar data, failing to effectively utilize forecast information for the forecast day (such as power levels at various times). This results in a relatively crude model construction and low assessment accuracy. Prior art document 2 constructs a probability density function using historical data, which results in fitting errors. It then samples data from this probability density function and calculates the net load probability density function based on typical scenarios derived from the sampled data. This sampling method also results in significant uncertainty and error. This makes it difficult to effectively ensure the accuracy of grid reserve capacity adjustment, and fails to effectively account for the impact of factors such as meteorological factors and power levels on forecast errors. Summary of the Invention

[0006] To address the deficiencies in the prior art, the present invention provides a method and system for evaluating the demand for operating reserve capacity that takes into account the uncertainty of renewable energy. This method and system consider the impact of factors such as the source and load sides, meteorological factors, and power size on prediction errors. It can dynamically evaluate the demand for operating reserve capacity (including upward and downward reserve) on the day before a high proportion of wind power is connected to the grid, thereby improving the accuracy of reserve demand assessment, thereby reducing the waste or shortage of reserve resources and achieving a balance between system reliability and economy.

[0007] The present invention adopts the following technical solutions.

[0008] A first aspect of the present invention provides a method for evaluating operating reserve capacity requirements taking into account the uncertainty of new energy sources, comprising:

[0009] Obtain historical data on new energy and load of the power grid, perform preprocessing and build new energy database and load database;

[0010] Calculate the new energy forecast error and load forecast error and their correlation with the corresponding influencing factors respectively, and construct the comprehensive influencing factors of the new energy forecast error and the load forecast error based on the corresponding correlations;

[0011] Based on the comprehensive influencing factors of new energy forecast errors and load forecast errors, the new energy database and load database are clustered and divided respectively to obtain multiple new energy sub-databases and multiple load sub-databases;

[0012] Construct the new energy forecast error probability distribution under each new energy sub-database and the load forecast error probability distribution under each load sub-database respectively. Obtain the net load forecast error value based on the new energy forecast error probability distribution and the load forecast error probability distribution. Use the convolution algorithm to integrate the net load forecast error value to obtain the net load forecast probability distribution.

[0013] Obtain the new energy forecast data and load forecast data for the planned day, and construct a time series of comprehensive influencing factors of the new energy forecast error and a time series of comprehensive influencing factors of the load forecast error for the planned day; calculate the similarity between the time series of comprehensive influencing factors of the new energy forecast error and the time series of comprehensive influencing factors of the load forecast error for the planned day and each new energy sub-database and load sub-database, select the new energy sub-database and load sub-database corresponding to the maximum similarity and determine them as the new energy sub-database and load sub-database for the planned day, thereby obtaining the corresponding net load forecast probability distribution;

[0014] The operating reserve capacity requirement is calculated based on the net load forecast probability distribution and the preset confidence level.

[0015] Optionally, the influencing factors corresponding to the new energy prediction error include at least one of the following: regional average temperature, regional average wind speed, daily average predicted power, and new energy power prediction time;

[0016] The influencing factors corresponding to the load forecast error include at least one of the following: daily average load forecast power, peak and valley electricity price periods, and load forecast time.

[0017] Optionally, the new energy historical data includes new energy power forecast data and new energy power measured data, and the load historical data includes load forecast data and load measured data. The new energy forecast error and the load forecast error, as well as their correlations with corresponding influencing factors, are calculated respectively, and a comprehensive influencing factor of the new energy forecast error and a comprehensive influencing factor of the load forecast error are constructed based on the corresponding correlations, including:

[0018] Calculate the difference between the new energy power prediction data and the new energy power measured data to obtain the new energy prediction error; calculate the difference between the load power prediction data and the load power measured data to obtain the load prediction error;

[0019] Normalize each influencing factor and map it to a unified interval;

[0020] The Pearson correlation coefficient method is used to calculate the correlation between the new energy forecast error and the load forecast error and the corresponding normalized influencing factors, and the correlation coefficient between the new energy forecast error and the corresponding influencing factors and the correlation coefficient between the load forecast error and the corresponding influencing factors are obtained;

[0021] The correlation coefficients of each influencing factor corresponding to the new energy forecast error are weighted and summed to obtain the comprehensive influencing factor of the new energy forecast error. The correlation coefficients of each influencing factor corresponding to the load forecast error are weighted and summed to obtain the comprehensive influencing factor of the load forecast error.

[0022] Optionally, the new energy prediction error and the weight of the i-th influencing factor are obtained according to the following formula:

[0023]

[0024] Where, Indicates the i-th influencing factor The weight of represents the i-th influencing factor, Represents the new energy prediction error and influencing factors The correlation coefficient of .

[0025] Optionally, based on the comprehensive impact factor of the new energy forecast error and the comprehensive impact factor of the load forecast error, the new energy database and the load database are clustered and divided respectively to obtain multiple new energy sub-databases and multiple load sub-databases, including:

[0026] Add the comprehensive impact factor of new energy prediction error to the new energy database;

[0027] Add the comprehensive impact factors of load forecast error to the load database;

[0028] Based on the comprehensive impact factor of new energy prediction error, the optimal number of clusters and cluster centers are found in the new energy database. Based on the optimal clusters and cluster centers, the new energy database is clustered and divided to form multiple new energy sub-databases.

[0029] Based on the comprehensive influencing factors of load forecast error, the optimal number of clusters and cluster centers are found in the load database. Based on the optimal clusters and cluster centers, the load database is clustered and divided to form multiple load sub-databases.

[0030] Optionally, the probability distribution of new energy prediction errors under each new energy sub-database is constructed according to the following formula:

[0031]

[0032] Where, is the probability density function of new energy prediction error, is the new energy prediction error, is the i-th prediction error value corresponding to the new energy sub-database, is the number of new energy prediction error samples in the new energy sub-database, is the bandwidth parameter of the kernel function, is the kernel function.

[0033] Optionally, the optimal bandwidth parameter is determined based on the mean integrated square error of the new energy prediction error.

[0034] Optionally, the optimal bandwidth parameter is determined based on the mean integrated square error of the new energy prediction error, including:

[0035] Based on experience, multiple bandwidth parameters are selected and combined with new energy prediction error samples to calculate multiple corresponding new energy prediction error probability distributions;

[0036] The difference between the probability distribution of the prediction error of each new energy source and the probability distribution of the prediction error in the corresponding new energy real sample is calculated, and the square of the difference is integrated and averaged to obtain the average integrated square error value corresponding to each bandwidth parameter;

[0037] The bandwidth parameter corresponding to the minimum mean integrated square error value is selected as the optimal bandwidth parameter.

[0038] Optionally, the similarity between the time series of the comprehensive influencing factors of the planned day new energy forecast error and the time series of the comprehensive influencing factors of the planned day load forecast error and each new energy sub-database and load sub-database is calculated separately, including:

[0039] Calculate the Euclidean distance between the time series of comprehensive influencing factors of new energy forecast errors on the planned day and the cluster center of each new energy sub-database as the degree of similarity;

[0040] The Euclidean distance between the time series of the comprehensive influencing factors of the planned day load forecast error and the cluster center of each load sub-database is calculated as the degree of similarity.

[0041] Optionally, the operating reserve capacity requirement is calculated based on the net load probability distribution and a preset confidence level as follows:

[0042]

[0043] Where, To meet the confidence level The operating reserve capacity requirement, is the net load probability distribution, The preset confidence level for the operational capacity reserve requirement, To meet the confidence level The next spare, To meet the confidence level The upper spare, for The lower quantile of the normal distribution, for The upper quantile of the normal distribution.

[0044] A second aspect of the present invention provides an operating spare capacity demand assessment system, the system comprising:

[0045] The first calculation module is used to calculate the new energy forecast error and the load forecast error and their correlations with the corresponding influencing factors, and construct a comprehensive influencing factor of the new energy forecast error and a comprehensive influencing factor of the load forecast error based on the corresponding correlations;

[0046] A clustering module is used to cluster the new energy database and the load database based on the comprehensive impact factor of the new energy forecast error and the comprehensive impact factor of the load forecast error, respectively, to obtain multiple new energy sub-databases and multiple load sub-databases;

[0047] The second calculation module is used to construct the new energy prediction error probability distribution under each new energy sub-database and the load prediction error probability distribution under each load sub-database, obtain the net load prediction error value based on the new energy prediction error probability distribution and the load prediction error probability distribution, and integrate the net load prediction error value using a convolution algorithm to obtain the net load prediction probability distribution;

[0048] The third calculation module is used to obtain the new energy forecast data and load forecast data for the planned day, and construct a time series of comprehensive influencing factors of the new energy forecast error and a time series of comprehensive influencing factors of the load forecast error for the planned day; respectively calculate the similarity between the time series of comprehensive influencing factors of the new energy forecast error and the time series of comprehensive influencing factors of the load forecast error for the planned day and each new energy sub-database and load sub-database, select the new energy sub-database and load sub-database corresponding to the maximum similarity and determine them as the new energy sub-database and load sub-database for the planned day, thereby obtaining the corresponding net load forecast probability distribution;

[0049] The fourth calculation module is used to calculate the operating reserve capacity demand based on the net load forecast probability distribution and the preset confidence level.

[0050] The third aspect of the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the computer program is loaded into the processor, the method for evaluating the demand for operating reserve capacity taking into account the uncertainty of new energy is implemented.

[0051] A fourth aspect of the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-mentioned method for evaluating the demand for operating reserve capacity taking into account the uncertainty of new energy.

[0052] Compared with the prior art, the beneficial effects of the present invention include at least:

[0053] This invention constructs comprehensive influencing factors to effectively account for the impact of meteorological factors, power level, and other factors on prediction errors. This factor, used to obtain the corresponding probability density function for each sub-database (sample set), results in a more accurate model. Furthermore, the data used is historically accurate, eliminating sampling errors and ensuring model accuracy. Furthermore, the invention obtains a probability density function for new energy / load based on each sub-database and, combined with a convolution algorithm, obtains a probability density function under net load prediction errors, thereby completing the backup capacity assessment. This invention effectively considers intraday operating information (differentiating between power level and meteorological conditions), resulting in a more accurate probability density function. Furthermore, by accounting for the impact of meteorological factors, power level, and other factors on prediction errors, it is more accurate than clustering power data alone. Furthermore, this invention considers the impact of uncertainty on both the source and load sides on backup capacity, making it more compliant with the current system backup requirements under high renewable energy penetration. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort. Among them:

[0055] Figure 1 This is a flow chart of a method for evaluating operating reserve capacity demand taking into account the uncertainty of new energy sources, provided by an embodiment of the present invention;

[0056] Figure 2 This is a schematic diagram of an improved hierarchical clustering-k-means algorithm flow provided by an embodiment of the present invention;

[0057] Figure 3 The figure is a flow chart of a method for determining a bandwidth parameter of a kernel function provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0058] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. The embodiments described in this application are only part of the embodiments of the present invention, not all of them. Based on the spirit of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0059] Combine Figure 1 As shown, embodiment 1 of the present invention provides a method for evaluating the demand for operating reserve capacity taking into account the uncertainty of new energy sources, which specifically includes the following contents:

[0060] S1: Obtain historical data on renewable energy and load of the power grid, perform preprocessing and build a new energy database and a load database.

[0061] New energy historical data includes new energy power forecast data and new energy power measured data, and load historical data includes load forecast data and load measured data.

[0062] Optionally, preprocessing includes missing value processing, duplicate data removal, outlier processing, normalization, etc.

[0063] In this way, by obtaining historical data to provide data support for subsequent predictions, preprocessing the historical data can improve the accuracy of subsequent prediction data, thereby improving the accuracy of spare capacity demand prediction.

[0064] S2: Based on the historical data of new energy and load in S1, the correlation between the new energy forecast error and the load forecast error and each influencing factor is calculated respectively, and the comprehensive influencing factor of the new energy forecast error and the comprehensive influencing factor of the load forecast error are constructed based on the corresponding correlation.

[0065] The influencing factors corresponding to the new energy prediction error include at least one of the following: regional average temperature, regional average wind speed, daily average predicted power, and new energy power prediction time.

[0066] The influencing factors corresponding to the load forecast error include at least one of the following: daily average load forecast power, peak and valley electricity price periods, and load forecast time.

[0067] Specifically, the new energy prediction error and power prediction value , regional average temperature , regional average wind speed , daily average predicted power , energy power forecast time (for example, including energy power forecast time , New energy power forecast month Based on the Pearson / Spearman correlation coefficient or other correlation coefficients (Pearson correlation, Spearman correlation), the correlation between the new energy prediction error and each influencing factor is obtained.

[0068] Specifically, the load forecast error and load forecast value , daily average load forecast power , peak and valley electricity price periods , load forecast time (e.g. including load forecast month ) and other influencing factors. Based on the Pearson / Spearman correlation coefficient or other correlation coefficients (Pearson correlation, Spearman correlation), the correlation between the load forecast error and each influencing factor is obtained.

[0069] S2 specifically includes:

[0070] S2.1: Calculate the new energy forecast error and load forecast error based on the new energy historical data and load historical data respectively.

[0071] The new energy forecast error and load forecast error are calculated as follows:

[0072] , (1)

[0073] , (2)

[0074] Where, represents the new energy prediction error at time t, Indicates the actual value of new energy power at time t, represents the predicted value of new energy power at time t, represents the load forecast error at time t, represents the actual value of load power at time t, Represents the load power forecast value at time t.

[0075] S2.2: Calculate the correlation between the new energy forecast error and the load forecast error and the corresponding influencing factors respectively, and construct the comprehensive influencing factors of the new energy forecast error and the load forecast error based on the corresponding correlations. S2.2 includes:

[0076] S2.2.1: Normalize each influencing factor and use the Pearson correlation coefficient to measure the normalized influencing factor X ( )and Correlation of new energy forecast errors.

[0077] The normalization method may be a minimum-maximum normalization method, which uniformly maps each influencing factor to a preset interval, such as the interval [0, 1].

[0078] Specifically, the correlation between each influencing factor and the new energy prediction error is calculated as follows:

[0079] , (3)

[0080] Where, Represents the new energy prediction error and influencing factors Correlation coefficient of represents the new energy prediction deviation under the i-th influencing factor, represents the new energy forecast mean, represents the number of samples, represents the i-th influencing factor, Represents the mean of the influencing factors.

[0081] S2.2.2: Based on the normalized influencing factors and the correlation between each influencing factor and the new energy prediction error, construct a comprehensive influencing factor for the new energy prediction error.

[0082] Specifically, the comprehensive impact factor of new energy prediction error is constructed according to the following formula:

[0083] , (4)

[0084] , (5)

[0085] Where, represents the comprehensive impact factor of new energy prediction error, represents the weight of the i-th influencing factor, represents the i-th influencing factor, Represents the new energy prediction error and influencing factors The correlation coefficient of .

[0086] S2.2.3: Normalize each influencing factor and use the Pearson correlation coefficient to measure each normalized influencing factor. ( ) and load forecast error correlation.

[0087] It is understandable that the technical means of S2.2.3 are similar to those of S2.2.1 and will not be repeated here.

[0088] S2.2.4: Based on the normalized influencing factors and the correlation between each influencing factor and the load forecast error, construct a comprehensive influencing factor of the load forecast error.

[0089] Specifically, the comprehensive impact factor of load forecast error is constructed according to the following formula:

[0090] (6)

[0091] (7)

[0092] Where, represents the comprehensive impact factor of load forecast error, represents the weight of the i-th influencing factor, represents the i-th influencing factor, represents the load forecast error and influencing factors The correlation coefficient of .

[0093] In this embodiment, by considering the impact of multiple influencing factors on the prediction error, the demand for day-ahead operating reserve capacity (including upward reserve and downward reserve) after a high proportion of wind power is connected to the grid can be dynamically evaluated, thereby improving the accuracy of the reserve demand assessment.

[0094] S3: Based on the comprehensive impact factors of new energy forecast errors and load forecast errors, cluster the new energy database and load database respectively to obtain multiple new energy sub-databases and multiple load sub-databases. S3 specifically includes:

[0095] Add the comprehensive impact factor of new energy prediction error to the new energy database;

[0096] Add the comprehensive impact factors of load forecast error to the load database;

[0097] Based on the comprehensive impact factor of new energy prediction error, the optimal number of clusters and cluster centers are found in the new energy database. Based on the optimal clusters and cluster centers, the new energy database is clustered and divided to form multiple new energy sub-databases.

[0098] Based on the comprehensive influencing factors of load forecast error, the optimal number of clusters and cluster centers are found in the load database. Based on the optimal clusters and cluster centers, the load database is clustered and divided to form multiple load sub-databases.

[0099] The new energy database includes sequences such as new energy forecast data, new energy measured data, comprehensive influencing factors of new energy forecast errors, and new energy error data. The load database includes sequences such as load forecast data, load measured data, comprehensive influencing factors of load forecast errors, and load error data.

[0100] In some embodiments, the clustering algorithm may be a k-means algorithm, a hierarchical clustering algorithm, a DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm, or a Gaussian mixture model algorithm, which is not further limited in the present invention.

[0101] Combine Figure 2 As shown, the embodiment of the present invention uses an improved hierarchical k-means clustering algorithm to cluster the new energy database and load database.

[0102] Specifically, the new energy historical data and load historical data are divided by day (24 hours, 96 point data), and 96 comprehensive impact factors of new energy forecast errors are calculated. and comprehensive influencing factors of load forecast error , respectively added to the corresponding new energy database and load database, first set the number of clusters k in the new new energy database and load database, and randomly select k data points as the initial cluster center, then divide each sub-data set based on the Euclidean distance, use the mean method to re-correct the cluster center, and use the mean of each sub-data set as the new cluster center to determine whether it has converged (the change in the cluster center is less than the preset threshold or the sum of the squares of the distances from each point in the cluster to the cluster center is less than the preset change threshold for continuous iterations, then it is considered to have converged). If not, re-cluster to find the optimal number of clusters and cluster centers. Then calculate whether each cluster sample is less than the preset number threshold (for example, 10% of the total number of samples can be taken). If not, end the clustering, otherwise re-perform K-1 divisions and perform a new round of clustering iterations until the iteration stop condition is met. Comprehensive impact factor of new energy prediction error based on the optimal number of clusters and cluster centers and comprehensive influencing factors of load forecast error Clustering is performed to form multiple new energy sub-databases and multiple load sub-databases with significant differences.

[0103] In this way, by improving the hierarchical k-means clustering algorithm, the influence of the initial cluster center on the stability of the K-means clustering algorithm can be reduced, and by dynamically adjusting the k value, the efficiency and effect of clustering can be improved.

[0104] S4: Construct the new energy prediction error probability distribution under each new energy sub-database and the load prediction error probability distribution under each load sub-database respectively, obtain the net load prediction error value according to the new energy prediction error probability distribution and the load prediction error probability distribution, and use the convolution algorithm to integrate the net load prediction error value to obtain the net load prediction probability distribution.

[0105] Specifically, based on N renewable energy forecast errors and M load forecast errors, we generate N*M net load forecast error probability density functions. Since net load errors and backups have a corresponding relationship, we synthesize the net load probability density function through convolution. Clustering multiple sub-databases reduces sample requirements. When the sample size is small, a load interval in the original sample may not contain data from all renewable energy types. Few intervals can lead to insufficient accuracy.

[0106] S4 specifically includes:

[0107] S4.1: Use non-parametric estimation methods to construct the probability distribution of new energy prediction errors under each sub-database and load forecast error probability distribution .

[0108] S4.1.1: Obtain the probability distribution of new energy forecast errors according to the following formula:

[0109] (8)

[0110] Where, is the new energy prediction error, is the probability density function of new energy prediction error (i.e., the probability distribution of new energy prediction error), is the i-th prediction error value corresponding to the new energy sub-database, is the number of new energy prediction error samples in the new energy sub-database, is the bandwidth parameter of the kernel function, is the kernel function.

[0111] It is understandable that the kernel function may be a Gaussian kernel function, a uniform kernel function, a triangular kernel function, etc.

[0112] It is understandable that bandwidth The choice of has a great influence on the kernel density estimation. A too wide bandwidth can easily lead to bias in the estimation, while a too narrow bandwidth can easily lead to noise in the estimation, which in turn leads to inaccurate prediction of the spare capacity demand.

[0113] Optionally, calculate the bandwidth parameter as follows:

[0114] (9)

[0115] Where c is the dimensionless coefficient, is the standard deviation of the new energy forecast error sample, is the number of new energy prediction error samples in the new energy sub-database.

[0116] Optionally, the optimal bandwidth parameter is determined based on the mean integrated square error of the new energy prediction error.

[0117] Optional, combined Figure 3 , the optimal bandwidth parameters are determined based on the mean integrated square error of the new energy prediction error, including:

[0118] Based on experience, multiple bandwidth parameters are selected and combined with new energy prediction error samples to calculate multiple corresponding new energy prediction error probability distributions;

[0119] The difference between the probability distribution of the prediction error of each new energy source and the probability distribution of the prediction error in the corresponding new energy real sample is calculated, and the square of the difference is integrated and averaged to obtain the average integrated square error value corresponding to each bandwidth parameter;

[0120] The bandwidth parameter corresponding to the minimum mean integrated square error value is selected as the optimal bandwidth parameter.

[0121] Specifically, the difference between the prediction error probability distribution of each new energy source and the prediction error probability distribution of the corresponding new energy real sample is calculated according to the following formula, and the square of the difference is integrated and averaged to obtain the average integrated square error value corresponding to each bandwidth parameter:

[0122] (10)

[0123] Where, for The corresponding mean integrated square error value is, For the desired, in this embodiment, it can be an average value, is the probability density function of new energy prediction error, is the prediction error in the real sample Probability density.

[0124] Specifically, determine the number N of new energy prediction errors in the new energy sub-database, obtain a set of bandwidth parameters according to formula (9), and then divide the new energy prediction error samples in the new energy data word library into p parts. Each time, select p-1 parts for training according to formula (10). The training results are averaged to obtain the prediction error probability distribution, and the remaining 1 part is selected for testing to obtain the prediction error probability in the real sample. ,choose The h corresponding to the minimum value is taken as the optimal bandwidth value.

[0125] S4.1.2: Obtain the load forecast error probability distribution according to the following formula:

[0126] (11)

[0127] Where, is the load forecast error, is the probability density function of load forecast error, is the i-th prediction error value corresponding to the load forecast sub-database, is the number of load forecast errors in the load forecast sub-database, is the bandwidth parameter of the kernel function, is the kernel function.

[0128] It is understandable that the bandwidth parameter of the kernel function of the load forecast error probability distribution can be set in the same way as the bandwidth parameter of the kernel function of the load forecast error probability distribution, which will not be described in detail.

[0129] S4.2: Calculate the net load forecast error value based on the probability distribution of the new energy forecast error and the load forecast error in each sub-database. Use the convolution algorithm to integrate the net load forecast error value to obtain the net load forecast probability distribution. This effectively considers the impact of factors such as meteorological factors and power level on the forecast error, and uses this to calculate the corresponding probability density function for each sub-database (sample set). This model is more accurate, and the data used is real historical data, eliminating sampling errors and ensuring model accuracy.

[0130] Specifically, the net load forecast probability distribution is calculated according to the following formula:

[0131] (12)

[0132] (13)

[0133] Where, is the net load forecast probability distribution, is the net load forecast error value, is the load forecast error, is the new energy prediction error.

[0134] In this embodiment, the new energy / load probability density function is obtained based on each sub-database, and combined with the convolution algorithm, the probability density function under the net load prediction error is obtained to complete the backup capacity evaluation.

[0135] S5: Obtain the new energy forecast data and load forecast data for the planned day, and construct a time series of comprehensive influencing factors of the new energy forecast error and a time series of comprehensive influencing factors of the load forecast error for the planned day; calculate the similarity between the time series of comprehensive influencing factors of the new energy forecast error and the time series of comprehensive influencing factors of the load forecast error for the planned day and each new energy sub-database and load sub-database respectively, select the new energy sub-database and load sub-database corresponding to the maximum similarity as the new energy sub-database and load sub-database for the planned day, and thereby obtain the corresponding net load forecast probability distribution.

[0136] Optionally, the similarity between the time series of the comprehensive influencing factors of the planned day new energy forecast error and the time series of the comprehensive influencing factors of the planned day load forecast error and each new energy sub-database and load sub-database is calculated separately, including:

[0137] Calculate the Euclidean distance between the time series of comprehensive influencing factors of new energy forecast errors on the planned day and the cluster center of each new energy sub-database as the degree of similarity;

[0138] The Euclidean distance between the time series of the comprehensive influencing factors of the planned day load forecast error and the cluster center of each load sub-database is calculated as the degree of similarity.

[0139] Specifically, the Euclidean distance from the time series of comprehensive influencing factors of new energy forecast errors on planned days to the cluster center of each new energy sub-database is calculated as follows:

[0140] (14)

[0141] Where, is the comprehensive impact factor sequence of new energy prediction error of the cluster center of each new energy sub-database, is the time series of comprehensive influencing factors of new energy forecast error on the planned day, This is the Euclidean distance from the time series of the comprehensive influencing factors of the new energy forecast error on the planned day to the cluster center of each new energy sub-database. A smaller Euclidean distance indicates a greater degree of similarity. The sub-database to which the planned day belongs is determined based on this Euclidean distance.

[0142] It should be noted that the load category confirmation is carried out in the same manner, which will not be described in detail. After determining the sub-database in which it is located, the net load prediction probability distribution is obtained according to S3.

[0143] S6: Calculate the operating reserve capacity requirement based on the net load probability distribution and the preset confidence level.

[0144] Specifically, if the confidence level is , then the backup demand that meets a certain confidence level can be expressed as:

[0145] (15)

[0146] Where, To meet the confidence level The operating reserve capacity requirement, is the net load probability distribution, The preset confidence level for the operational capacity reserve requirement, To meet the confidence level The next spare, To meet the confidence level The upper spare, for The lower quantile of the normal distribution, for The upper quantile of the normal distribution, hour, , , 、 for The possible values ​​of .

[0147] Embodiment 2 of the present invention provides an operating reserve capacity demand assessment system, which implements the operating reserve capacity demand assessment method taking into account the uncertainty of new energy sources as described in embodiment 1. The system includes:

[0148] The first calculation module is used to calculate the new energy forecast error and the load forecast error and their correlations with the corresponding influencing factors, and construct a comprehensive influencing factor of the new energy forecast error and a comprehensive influencing factor of the load forecast error based on the corresponding correlations;

[0149] A clustering module is used to cluster the new energy database and the load database based on the comprehensive impact factor of the new energy forecast error and the comprehensive impact factor of the load forecast error, respectively, to obtain multiple new energy sub-databases and multiple load sub-databases;

[0150] The second calculation module is used to construct the new energy prediction error probability distribution under each new energy sub-database and the load prediction error probability distribution under each load sub-database, obtain the net load prediction error value based on the new energy prediction error probability distribution and the load prediction error probability distribution, and integrate the net load prediction error value using a convolution algorithm to obtain the net load prediction probability distribution;

[0151] The third calculation module is used to obtain the new energy forecast data and load forecast data for the planned day, and construct a time series of comprehensive influencing factors of the new energy forecast error and a time series of comprehensive influencing factors of the load forecast error for the planned day; respectively calculate the similarity between the time series of comprehensive influencing factors of the new energy forecast error and the time series of comprehensive influencing factors of the load forecast error for the planned day and each new energy sub-database and load sub-database, select the new energy sub-database and load sub-database corresponding to the maximum similarity and determine them as the new energy sub-database and load sub-database for the planned day, thereby obtaining the corresponding net load forecast probability distribution;

[0152] The fourth calculation module is used to calculate the operating reserve capacity demand based on the net load forecast probability distribution and the preset confidence level.

[0153] Regarding the system in the above embodiment, the specific manner in which each unit performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0154] Embodiment 3 of the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the computer program is loaded into the processor, the method for evaluating the operating reserve capacity demand taking into account the uncertainty of new energy as described in embodiment 1 is implemented.

[0155] Embodiment 4 of the present invention provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the method for evaluating the operating reserve capacity demand taking into account the uncertainty of new energy according to embodiment 1 is implemented.

[0156] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0157] The present disclosure may be a system, method and / or computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for causing a processor to implement various aspects of the present disclosure.

[0158] A computer-readable storage medium can be a tangible device that can hold and store instructions for use by an instruction execution device. A computer-readable storage medium can be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), a portable compact disc read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanical encoding device, such as a punched card or raised structure in a groove on which instructions are stored, and any suitable combination thereof. As used herein, a computer-readable storage medium is not to be construed as a transient signal per se, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagating through a waveguide or other transmission medium (e.g., a light pulse passing through a fiber optic cable), or an electrical signal transmitted through an electrical wire.

[0159] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to each computing / processing device, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network can include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. The network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions to be stored in the computer-readable storage medium in each computing / processing device.

[0160] The computer program instructions for performing the operations of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The computer-readable program instructions may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, the state information of the computer-readable program instructions is used to personalize an electronic circuit, such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), so that the electronic circuit can execute the computer-readable program instructions, thereby implementing various aspects of the present disclosure.

[0161] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the scope of protection of the claims of the present invention.

Claims

1. A method for evaluating the demand for operating reserve capacity taking into account the uncertainty of new energy sources, characterized in that: include: Obtain historical data on new energy and load of the power grid, perform preprocessing and build new energy database and load database; Calculate the new energy forecast error and load forecast error and their correlation with the corresponding influencing factors respectively, and construct the comprehensive influencing factors of the new energy forecast error and the load forecast error based on the corresponding correlations; Based on the comprehensive influencing factors of new energy forecast errors and load forecast errors, the new energy database and load database are clustered and divided respectively to obtain multiple new energy sub-databases and multiple load sub-databases; Construct the new energy forecast error probability distribution under each new energy sub-database and the load forecast error probability distribution under each load sub-database respectively. Obtain the net load forecast error value based on the new energy forecast error probability distribution and the load forecast error probability distribution. Use the convolution algorithm to integrate the net load forecast error value to obtain the net load forecast probability distribution. Obtain the new energy forecast data and load forecast data for the planned day, and construct a time series of comprehensive influencing factors of the new energy forecast error and a time series of comprehensive influencing factors of the load forecast error for the planned day; calculate the similarity between the time series of comprehensive influencing factors of the new energy forecast error and the time series of comprehensive influencing factors of the load forecast error for the planned day and each new energy sub-database and load sub-database, select the new energy sub-database and load sub-database corresponding to the maximum similarity and determine them as the new energy sub-database and load sub-database for the planned day, thereby obtaining the corresponding net load forecast probability distribution; Calculate the operating reserve capacity requirement based on the net load forecast probability distribution and the preset confidence level; The new energy historical data includes new energy power forecast data and new energy power measured data, and the load historical data includes load forecast data and load measured data. The new energy forecast error and load forecast error and their correlation with the corresponding influencing factors are calculated respectively, and the comprehensive influencing factors of the new energy forecast error and the comprehensive influencing factors of the load forecast error are constructed based on the corresponding correlations, including: Calculate the difference between the new energy power prediction data and the new energy power measured data to obtain the new energy prediction error; calculate the difference between the load power prediction data and the load power measured data to obtain the load prediction error; Normalize each influencing factor and map it to a unified interval; The Pearson correlation coefficient method is used to calculate the correlation between the new energy forecast error and the load forecast error and the corresponding normalized influencing factors, and the correlation coefficient between the new energy forecast error and the corresponding influencing factors and the correlation coefficient between the load forecast error and the corresponding influencing factors are obtained; The correlation coefficients of each influencing factor corresponding to the new energy forecast error are weighted and summed to obtain the comprehensive influencing factor of the new energy forecast error. The correlation coefficients of each influencing factor corresponding to the load forecast error are weighted and summed to obtain the comprehensive influencing factor of the load forecast error. Based on the comprehensive influencing factors of new energy forecast errors and load forecast errors, the new energy database and load database are clustered and divided respectively, and multiple new energy sub-databases and multiple load sub-databases are obtained, including: Add the comprehensive impact factor of new energy prediction error to the new energy database; Add the comprehensive impact factors of load forecast error to the load database; Based on the comprehensive impact factor of new energy prediction error, the optimal number of clusters and cluster centers are found in the new energy database. Based on the optimal clusters and cluster centers, the new energy database is clustered and divided to form multiple new energy sub-databases. Based on the comprehensive influencing factors of load forecast error, the optimal number of clusters and cluster centers are found in the load database. Based on the optimal clusters and cluster centers, the load database is clustered and divided to form multiple load sub-databases. The operating reserve capacity requirement is calculated based on the net load probability distribution and the preset confidence level as follows: Where, To meet the confidence level The operating reserve capacity requirement, is the net load probability distribution, The preset confidence level for the operational capacity reserve requirement, To meet the confidence level The next spare, To meet the confidence level The upper spare, for The lower quantile of the normal distribution, for The upper quantile of the normal distribution.

2. The method for evaluating operating reserve capacity demand taking into account the uncertainty of new energy sources according to claim 1, characterized in that: The influencing factors corresponding to the new energy prediction error include at least one of the following: regional average temperature, regional average wind speed, daily average predicted power, and new energy power prediction time; The influencing factors corresponding to the load forecast error include at least one of the following: daily average load forecast power, peak and valley electricity price periods, and load forecast time.

3. The method for evaluating operating reserve capacity demand considering the uncertainty of new energy sources according to claim 1, characterized in that: The new energy prediction error and the weight of the i-th influencing factor are obtained according to the following formula: Where, Indicates the i-th influencing factor The weight of represents the i-th influencing factor, Represents the new energy prediction error and influencing factors The correlation coefficient of .

4. The method for evaluating operating reserve capacity demand considering uncertainty of new energy sources according to claim 1, characterized in that: The probability distribution of new energy prediction errors under each new energy sub-database is constructed according to the following formula: Where, is the probability density function of new energy prediction error, is the new energy prediction error, is the i-th prediction error value corresponding to the new energy sub-database, is the number of new energy prediction error samples in the new energy sub-database, is the bandwidth parameter of the kernel function, is the kernel function.

5. The method for evaluating operating reserve capacity demand taking into account the uncertainty of new energy sources according to claim 4 is characterized in that: The optimal bandwidth parameter is determined based on the mean integrated square error of the new energy prediction error.

6. The method for evaluating operating reserve capacity demand taking into account the uncertainty of new energy sources according to claim 5, characterized in that: The optimal bandwidth parameters are determined based on the mean integrated square error of the new energy prediction error, including: Based on experience, multiple bandwidth parameters are selected and combined with new energy prediction error samples to calculate multiple corresponding new energy prediction error probability distributions; The difference between the probability distribution of the prediction error of each new energy source and the probability distribution of the prediction error in the corresponding new energy real sample is calculated, and the square of the difference is integrated and averaged to obtain the average integrated square error value corresponding to each bandwidth parameter; The bandwidth parameter corresponding to the minimum mean integrated square error value is selected as the optimal bandwidth parameter.

7. The method for evaluating operating reserve capacity demand considering uncertainty of new energy sources according to claim 1, characterized in that: The similarity between the time series of comprehensive influencing factors of the planned day new energy forecast error and the time series of comprehensive influencing factors of the planned day load forecast error and each new energy sub-database and load sub-database is calculated separately, including: Calculate the Euclidean distance between the time series of comprehensive influencing factors of new energy forecast errors on the planned day and the cluster center of each new energy sub-database as the degree of similarity; The Euclidean distance between the time series of the comprehensive influencing factors of the planned day load forecast error and the cluster center of each load sub-database is calculated as the degree of similarity.

8. An operating reserve capacity demand assessment system using the operating reserve capacity demand assessment method taking into account the uncertainty of new energy sources according to any one of claims 1 to 7, characterized in that: The system comprises: The acquisition module is used to obtain the historical data of new energy and load of the power grid, perform preprocessing and build the new energy database and load database; The first calculation module is used to calculate the new energy forecast error and the load forecast error and their correlations with the corresponding influencing factors, and construct a comprehensive influencing factor of the new energy forecast error and a comprehensive influencing factor of the load forecast error based on the corresponding correlations; A clustering module is used to cluster the new energy database and the load database based on the comprehensive impact factor of the new energy forecast error and the comprehensive impact factor of the load forecast error, respectively, to obtain multiple new energy sub-databases and multiple load sub-databases; The second calculation module is used to construct the new energy prediction error probability distribution under each new energy sub-database and the load prediction error probability distribution under each load sub-database, obtain the net load prediction error value according to the new energy prediction error probability distribution and the load prediction error probability distribution, and integrate the net load prediction error value using a convolution algorithm to obtain the net load prediction probability distribution; The third calculation module is used to obtain the new energy forecast data and load forecast data for the planned day, and construct a time series of comprehensive influencing factors of the new energy forecast error and a time series of comprehensive influencing factors of the load forecast error for the planned day; respectively calculate the similarity between the time series of comprehensive influencing factors of the new energy forecast error and the time series of comprehensive influencing factors of the load forecast error for the planned day and each new energy sub-database and load sub-database, select the new energy sub-database and load sub-database corresponding to the maximum similarity and determine them as the new energy sub-database and load sub-database for the planned day, thereby obtaining the corresponding net load forecast probability distribution; The fourth calculation module is used to calculate the operating reserve capacity demand based on the net load forecast probability distribution and the preset confidence level.

9. An electronic device comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is configured to operate according to the instructions to execute the steps of the method for evaluating operating reserve capacity demand taking into account uncertainty of new energy sources according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method for evaluating the demand for operating reserve capacity taking into account the uncertainty of new energy sources as described in any one of claims 1 to 7 are implemented.

Images