A low-carbon optimal scheduling method based on source-network-load-storage integration

Abstract

The present invention discloses a low-carbon optimization scheduling method based on source-grid-load-storage integration, which relates to the technical field of energy scheduling and includes: obtaining historical power demand, renewable energy power generation, energy storage device status, and grid load data based on the environmental sensors of the source-grid-load-storage integration system; obtaining historical time-series characteristic parameter data of the operation mode of the source-grid-load-storage integration system, and determining the interval vector range of the operation mode of the source-grid-load-storage integration system; obtaining real-time data; obtaining a real-time operation vector matrix; performing a matching degree verification based on the real-time operation vector matrix and the interval vector range of the operation mode of the source-grid-load-storage integration system, and screening out the operation mode with the largest matching degree as the scheduling plan. The present invention optimizes the scheduling plan through the operation mode matching of source-grid-load-storage integration, and improves the utilization efficiency of grid load and energy storage devices.

Description

Technical Field

[0001] The present invention relates to the technical field of energy scheduling, and in particular, to a low-carbon optimal scheduling method based on the integration of power generation, grid, load and energy storage. Background Art

[0002] With the transformation of the energy structure and the increasing requirements for low-carbon environmental protection, traditional energy scheduling methods have been difficult to meet the changing power demand and the uncertainty of renewable energy generation. The integrated power generation, grid, load and energy storage system, that is, the coordinated operation of integrated power generation, load, energy storage equipment and the grid, plays an increasingly important role in the optimal scheduling of the power system. By real-time monitoring and data analysis of power demand, renewable energy generation, grid load and the state of energy storage equipment, the scheduling flexibility, efficiency and low-carbon performance of the power system can be improved. However, traditional scheduling methods usually rely on single historical data and lack in-depth mining of system operation modes and data correlation characteristics.

[0003] Therefore, how to combine the actual operation mode of the integrated power generation, grid, load and energy storage system, optimize the scheduling scheme based on the trend analysis of historical time-series data and real-time data, and reduce carbon emissions has become an important issue in the intelligent power system. Summary of the Invention

[0004] To solve the above technical problems, a low-carbon optimal scheduling method based on the integration of power generation, grid, load and energy storage is provided, and the technical solution solves the above problems.

[0005] To achieve the above object, the technical solution adopted by the present invention is as follows:[[]]

[0006] A low-carbon optimal scheduling method based on the integration of power generation, grid, load and energy storage, comprising:[[]]

[0007] Obtaining historical power demand, renewable energy generation, energy storage equipment status and grid load data based on environmental sensors of the integrated power generation, grid, load and energy storage system;

[0008] According to the operation mode of the integrated power generation, grid, load and energy storage system, screening the historical power demand, renewable energy generation, energy storage equipment status and grid load data to obtain historical time-series characteristic parameter data of the operation mode of the integrated power generation, grid, load and energy storage system;

[0009] Based on the historical time-series characteristic parameter data of the integrated power generation, grid, load and energy storage system, determining the range of interval vectors of the operation mode of the integrated power generation, grid, load and energy storage system;

[0010] Obtaining real-time power demand, renewable energy generation, energy storage equipment status and grid load data;

[0011] According to the operation mode of the source-grid-load-storage integrated system, perform correlation mapping on real-time power demand, renewable energy power generation, energy storage device status, and grid load data to obtain a real-time operation vector matrix;

[0012] Verify the matching degree according to the interval vector range of the real-time operation vector matrix and the operation mode of the source-grid-load-storage integrated system, and select the operation mode with the maximum matching degree as the dispatching plan.

[0013] Preferably, screen the historical power demand, renewable energy power generation, energy storage device status, and grid load data according to the operation mode of the source-grid-load-storage integrated system to obtain the historical time series characteristic parameter data of the source-grid-load-storage integrated system, specifically including:

[0014] Based on the historical power demand, renewable energy power generation, energy storage device status, and grid load data, calculate the correlation coefficient between each operation parameter and the operation mode of the source-grid-load-storage integrated system;

[0015] Cluster the correlation coefficients between each operation parameter and the operation mode of the source-grid-load-storage integrated system to obtain an array of correlation characteristic parameters of the operation mode of the source-grid-load-storage integrated system;

[0016] Based on the timestamps in the historical power demand, renewable energy power generation, energy storage device status, and grid load data, use the timestamp as the observation window and the array of correlation characteristic parameters of the operation mode of the source-grid-load-storage integrated system as the observation object to establish the historical time series characteristic parameter data of the source-grid-load-storage integrated system.

[0017] Preferably, the calculation of the correlation coefficient between each operation parameter and the operation mode of the source-grid-load-storage integrated system based on the historical power demand, renewable energy power generation, energy storage device status, and grid load data specifically includes:

[0018] Based on the operation mode of the source-grid-load-storage integrated system, mark the key operation parameters concerned by the system;

[0019] Assign a weighted mean to each key operation parameter;

[0020] Standardize the historical power demand, renewable energy power generation, energy storage device status, and grid load data, and construct a scatter plot of historical operation data;

[0021] Based on the scatter plot of historical operation data, key operation parameters, and weighted mean, calculate the correlation coefficient between each operation parameter and the key operation parameter in the historical operation data per unit time.

[0022] Preferably, the correlation coefficient is specifically: ; where is the The correlation coefficient between a key operating parameter and the th operating parameter, is the value of the th operating parameter, is the weighted mean of the key operating parameters, and

[0023] is the total number of operating parameters.

[0024] Preferably, cluster the correlation coefficients between each operating parameter and the operation mode of the source-grid-load-storage integrated system to obtain an array of correlation characteristic parameters of the system operation mode, specifically including:

[0025] Based on the correlation coefficients between each operating parameter and the source-grid-load-storage integrated system, establish a correlation coefficient matrix;

[0026] Perform iterative clustering through the k-means clustering algorithm: ; i represents the index of the i-th cluster, i = 1, 2,..., K;

[0027] Step a Initialization: Randomly select K cluster centers in the correlation coefficient matrix as the initial cluster centroids and the current cluster centroid ; calculate the Euclidean distance between each data point and the current cluster centroid, and assign each point to the nearest cluster to form a new cluster set

[0028] ; k represents the number of iterations, that is, the clustering algorithm is currently in the k-th round of iteration; ; is the number of data points in the i-th cluster;

[0029] Step d Iteration termination: Repeat steps b - c until the change in the centroid is less than the threshold or reaches the maximum number of iterations. At this time, the centroid is the array of correlation characteristic parameters of the system operation mode.

[0030] Preferably, based on the historical time series characteristic parameter data of the source-grid-load-storage integrated system, determine the interval vector range of the operation mode of the source-grid-load-storage integrated system, specifically including:

[0031] Substitute the historical time series data into the autoregressive moving average model to predict the operation mode in the future unit time, and determine the interval vector range through the confidence interval of the error function.

[0032] Preferably, the autoregressive moving average model is specifically: ; where is the predicted value at time, is the unit time, is the time offset, are all model coefficients, is the white noise error term; P and Q are the orders of autoregression and moving average respectively, which are determined by fitting historical data;

[0033] Interval vector range is determined by the predicted value and the confidence interval;

[0034] Among them, ; among them, is the quantile of the standard normal distribution corresponding to the confidence level α, and σ is the standard deviation of the prediction error.

[0035] Preferably, according to the real-time operation vector matrix and the interval vector range of the operation mode of the source-network-load-storage integrated system, the matching degree is verified, and the operation mode with the largest matching degree is selected, specifically including:

[0036] Based on the real-time operation vector matrix, calculate the score recursion value of each operation mode per unit time;

[0037] Verify the matching degree between the score recursion value and the interval vector range of the operation mode to obtain the matching degree value;

[0038] Based on the matching degree value, select the operation mode with the largest matching degree as the scheduling scheme.

[0039] Preferably, through the formula quantify the real-time operation vector matrix to obtain the real-time parameter value ; is the original value of the th parameter at time t; is the historical mean of the th parameter, which is used to centralize the data; is the historical standard deviation of the th parameter, which is used to scale the data distribution;

[0040] Calculate the score recursion value through the formula ; in the formula, the score recursion value represents the dynamic score of the th time unit and the th operation mode, reflecting the matching degree between the real-time data and the historical mode. The lower the score, the higher the matching degree; , are the upper and lower limits of the interval vector range respectively, is the real-time data fitting value, λ is the weight coefficient, 0 < λ < 1, which reflects the predicted value and the synergistic effect of the real-time parameter value​

[0041] When verifying the matching degree between the score recursion value and the interval vector range of the operation mode, if the real-time parameter falls within the interval vector range , the matching degree value is 1; otherwise it is 0;

[0042] Calculate the distance between the real-time parameter and the midpoint of the interval vector range , which is calculated through the formula ; The closer the value is to 1, the higher the matching degree.

[0043] Preferably, combining the score recursion value and the interval matching result, the final matching degree value is calculated by weighted calculation through the formula ; in the formula, is the interval matching weight;

[0044] Select the mode with the largest value as the scheduling scheme.

[0045] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0046] The present invention proposes to deeply analyze the changing trends of historical power demand, renewable energy power generation, energy storage device status, and grid load data. This method can accurately predict future power demand and power generation, provide a reliable basis for power dispatching, optimize the dispatching scheme through the operation mode matching of source-grid-load-storage integration, and improve the utilization efficiency of grid load and energy storage devices; at the same time, considering the uncertainty of renewable energy, through real-time monitoring and optimized dispatching strategies, improve the consumption capacity of renewable energy; in addition, low-carbon optimized dispatching can preferentially use green energy, reduce high carbon emissions, and promote the construction of a low-carbon power system. Based on the autoregressive moving average model and iterative clustering algorithm, the system dynamically adjusts the dispatching strategy to enhance the stability and flexibility of the power system and effectively cope with the risks brought by load fluctuations and energy storage imbalances. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 is a flowchart of a low-carbon optimized dispatching method based on source-grid-load-storage integration according to the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0048] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and other obvious variants can be thought of by those skilled in the art.

[0049] Referring to Figure 1 shown, a low-carbon optimized dispatching method based on source-grid-load-storage integration includes:

[0050] An environmental sensor based on the source-grid-load-storage integrated system acquires historical power demand, renewable energy power generation, energy storage device status, and grid load data;

[0051] According to the operation mode of the source-grid-load-storage integrated system, screen the historical power demand, renewable energy power generation, energy storage device status, and grid load data to obtain the historical time-series characteristic parameter data of the operation mode of the source-grid-load-storage integrated system;

[0052] Based on the historical time-series characteristic parameter data of the source-grid-load-storage integrated system, determine the interval vector range of the operation mode of the source-grid-load-storage integrated system;

[0053] Obtain real-time power demand, renewable energy power generation, energy storage device status, and grid load data;

[0054] According to the operation mode of the source-grid-load-storage integrated system, perform correlation mapping on the real-time power demand, renewable energy power generation, energy storage device status, and grid load data to obtain a real-time operation vector matrix;

[0055] According to the matching degree verification between the real-time operation vector matrix and the interval vector range of the operation mode of the source-grid-load-storage integrated system, screen out the operation mode with the largest matching degree as the scheduling scheme.

[0056] Screen the historical power demand, renewable energy power generation, energy storage device status, and grid load data according to the operation mode of the source-grid-load-storage integrated system to obtain the historical time-series characteristic parameter data of the source-grid-load-storage integrated system, specifically including:

[0057] Based on the historical power demand, renewable energy power generation, energy storage device status, and grid load data, calculate the correlation coefficient between each operation parameter and the operation mode of the source-grid-load-storage integrated system;

[0058] Cluster the correlation coefficients between each operation parameter and the operation mode of the source-grid-load-storage integrated system to obtain an array of correlation characteristic parameters of the operation mode of the source-grid-load-storage integrated system;

[0059] Based on the timestamps in the historical power demand, renewable energy power generation, energy storage device status, and grid load data, use the timestamp as the observation window and the array of correlation characteristic parameters of the operation mode of the source-grid-load-storage integrated system as the observation object to establish the historical time-series characteristic parameter data of the source-grid-load-storage integrated system;

[0060] By analyzing the changing trends of power demand, renewable energy generation, energy storage device status, and grid load, and combining historical time-series characteristic parameter data, the interval vector range of different operating modes is identified. Through feature extraction and trend analysis of time series, more accurate predictions are made for real-time operation, and the future operating mode is further predicted through the autoregressive moving average model.

[0061] Based on historical power demand, renewable energy generation, energy storage device status, and grid load data, calculate the correlation coefficients between each operating parameter and the operating mode of the source-network-load-storage integrated system, specifically including:

[0062] Based on the operating mode of the source-network-load-storage integrated system, mark the key operating parameters concerned by the system;

[0063] Assign a weighted mean to each key operating parameter;

[0064] Standardize the historical power demand, renewable energy generation, energy storage device status, and grid load data, and construct a scatter plot of historical operating data;

[0065] Based on the scatter plot of historical operating data, key operating parameters, and weighted means, calculate the correlation coefficients between each operating parameter and the key operating parameter in the historical operating data per unit time;

[0066] The specific correlation coefficient is: ; where is the correlation coefficient between the th key operating parameter and the th operating parameter, is the value of the th operating parameter, is the weighted mean of the key operating parameter, is the total number of operating parameters.

[0067] By calculating the correlation coefficients between each operating parameter and the system operating mode, and using the mean clustering algorithm to cluster the data, an array of correlation characteristic parameters of the system is obtained. This method improves the identification ability of complex operating modes and helps to select the most suitable scheduling scheme for different power demands and load changes.

[0068] Cluster the correlation coefficients between each operating parameter and the operating mode of the source-network-load-storage integrated system to obtain an array of correlation characteristic parameters of the system operating mode, specifically including:

[0069] Based on the correlation coefficients between each operating parameter and the source-network-load-storage integrated system, establish a correlation coefficient matrix,

[0070] Perform iterative clustering through the mean clustering algorithm, specifically including:

[0071] Step a Initialization: Randomly select K cluster centers in the correlation coefficient matrix as the initial cluster centroids ; i represents the index of the i-th cluster, i = 1, 2, …, K;

[0072] Step b Assign data points: Calculate the Euclidean distance between each data point and the current cluster centroid , and assign each point to the nearest cluster to form a new set of clusters ; k represents the number of iterations, that is, the clustering algorithm is currently in the k-th round of iteration;

[0073] Step c Update cluster centers: Recalculate the centroid of each cluster ; is the number of data points in the i-th cluster;

[0074] Step d Iteration termination: Repeat Step b - Step c until the change in the centroid is less than the threshold or the maximum number of iterations is reached. At this time, the centroid is the array of correlation feature parameters of the system operation mode.

[0075] Based on the historical time-series feature data of the source-network-load-storage integrated system, determine the interval vector range of the operation mode, specifically including:

[0076] Substitute the historical time-series data into the autoregressive moving average model to predict the operation mode in the future unit time, and determine the interval vector range through the confidence interval of the error function.

[0077] The autoregressive moving average model is specifically: ; In the formula, is the predicted value at time, is the unit time, is the time offset, are all model coefficients, is the white noise error term; P and Q are the orders of autoregression and moving average respectively, determined by fitting the historical data;

[0078] The interval vector range is determined by the predicted value and the confidence interval;

[0079] Among them, ; Among them, is the standard normal distribution quantile corresponding to the confidence level α, and σ is the standard deviation of the prediction error.

[0080] Verify the matching degree according to the real-time operation vector matrix and the interval vector range of the operation mode of the source-network-load-storage integrated system, and screen out the operation mode with the largest matching degree, specifically including:

[0081] Calculate the score recursion value of each operation mode per unit time based on the real-time operation vector matrix;

[0082] Verify the matching degree between the score recursion value and the interval vector range of the operation mode to obtain the matching degree value;

[0083] Based on the matching degree value, screen out the operation mode with the highest matching degree as the scheduling scheme;

[0084] Specifically, through the formula Quantize the real-time operation vector matrix to obtain the real-time parameter value ; is the original value of the th parameter at time t; is the historical mean of the th parameter, used to centralize the data; is the historical standard deviation of the th parameter, used to scale the data distribution;

[0085] Calculate the score recursion value through the formula ; In the formula, the score recursion value represents the dynamic score of the th operation mode at the th time unit, reflecting the matching degree between the real-time data and the historical mode. The lower the score, the higher the matching degree; are the upper and lower limits of the interval vector range respectively, , is the real-time data fitting value, λ is the weight coefficient, 0 < λ < 1, reflecting the collaborative influence of the predicted value and the real-time parameter value ;

[0086] When verifying the matching degree between the score recursion value and the interval vector range of the operation mode, if the real-time parameter falls within the interval vector range , the matching degree value is 1; otherwise it is 0;

[0087] Calculate the distance between the real-time parameter and the midpoint of the interval vector range, calculated through the formula ; The closer the value is to 1, the higher the matching degree.

[0088] Combining the score recursion value and the interval matching result, calculate the final matching degree value through the formula by weighted calculation ; In the formula, is the interval matching weight; it can take the value of 0.7;​

[0089] Select the largest pattern as the scheduling scheme.

[0090] By performing matching degree verification on the range vector of the real-time operation vector matrix and the historical operation pattern, the most suitable scheduling scheme is screened out. This process can significantly reduce the incorrect selection of the scheduling scheme and improve the system scheduling efficiency and grid stability based on the score recursion value and matching degree verification.

[0091] In summary, the advantages of the present invention are as follows:

[0092] By deeply analyzing the changing trends of data such as historical power demand, renewable energy power generation, energy storage device status, and grid load, it is possible to more accurately predict future power demand and power generation, thereby providing a more reliable basis for power scheduling;

[0093] Through the operation mode matching of source-grid-load-storage integration, it is possible to select the most suitable scheduling scheme for the current state according to the correlation between real-time operation data and historical data, and optimize the use of grid load and energy storage devices;

[0094] This method can comprehensively consider the uncertainty of renewable energy power generation, and effectively improve the consumption capacity of renewable energy through real-time monitoring and scheduling strategies;

[0095] Through low-carbon optimized scheduling, it is not only possible to optimize energy consumption, but also to achieve the priority scheduling of green energy, reduce the energy consumption with high carbon emissions, and promote the construction of a low-carbon power system;

[0096] Based on the autoregressive moving average model and the iterative clustering algorithm, the system can dynamically adjust the scheduling strategy, improve the stability and flexibility of the power system, and reduce the risks brought by load fluctuations and energy storage imbalances;

[0097] This method conducts in-depth analysis on the historical time series characteristic parameter data of the source-grid-load-storage integration system, combines real-time data, and performs dynamic optimized scheduling to ensure the efficient operation of the power system under different times and conditions;

[0098] On the basis of the traditional scheduling model, a calculation method of the correlation coefficient based on key operation parameters and weighted mean is introduced. Through cluster analysis and data mining, the correlation characteristic parameter array that can best represent the system operation characteristics is found, so as to accurately formulate the scheduling strategy;

[0099] By clustering the correlation characteristic parameters through the mean clustering algorithm, different operation modes can be accurately classified, making the power scheduling not only more refined, but also able to quickly respond according to the actual needs of the power system and optimize the energy configuration;

[0100] After obtaining the real-time operation vector matrix, the matching degree is verified in combination with the interval vector range of the source-network-load-storage integrated system, and the best-matched operation mode is selected. This method can ensure the real-time performance and accuracy of the scheduling scheme, thereby improving the scheduling efficiency of the system.

[0101] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. What is described in the above embodiments and the specification is only the principle of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection required by the present invention is defined by the appended claims and their equivalents.

Claims

1. A low-carbon optimal scheduling method based on source-network-load-storage integration, characterized in that Including: An environmental sensor based on the source-network-load-storage integrated system to obtain historical power demand, renewable energy generation, energy storage device status, and grid load data; According to the operation mode of the source-network-load-storage integrated system, screen the historical power demand, renewable energy generation, energy storage device status, and grid load data to obtain the historical time-series characteristic parameter data of the operation mode of the source-network-load-storage integrated system, specifically including: Based on the historical power demand, renewable energy generation, energy storage device status, and grid load data, calculate the correlation coefficient between each operation parameter and the operation mode of the source-network-load-storage integrated system; Cluster the correlation coefficients between each operation parameter and the operation mode of the source-network-load-storage integrated system to obtain an array of correlation characteristic parameters of the operation mode of the source-network-load-storage integrated system; Based on the timestamps in the historical power demand, renewable energy generation, energy storage device status, and grid load data, use the timestamp as the observation window and the array of correlation characteristic parameters of the operation mode of the source-network-load-storage integrated system as the observation object to establish the historical time-series characteristic parameter data of the source-network-load-storage integrated system; Based on the historical time-series characteristic parameter data of the source-network-load-storage integrated system, determine the interval vector range of the operation mode of the source-network-load-storage integrated system, specifically including: Substitute historical time-series data into the autoregressive moving average model to predict the operating mode in the future unit time, and determine the range of the interval vector through the confidence interval of the error function; the autoregressive moving average model is specifically: ; In the formula, is the predicted value at time is the unit time, is the time offset, are all model coefficients, is the white noise error term; P and Q are the orders of autoregression and moving average respectively, which are determined by fitting historical data; Interval vector range Determined by the predicted value and the confidence interval; Among them, ; among them, is the quantile of the standard normal distribution corresponding to the confidence level α, and σ is the standard deviation of the prediction error; Obtain real-time power demand, renewable energy generation, energy storage device status, and grid load data; According to the operation mode of the source-network-load-storage integrated system, perform correlation mapping on the real-time power demand, renewable energy generation, energy storage device status, and grid load data to obtain a real-time operation vector matrix; According to the matching degree verification between the real-time operation vector matrix and the interval vector range of the operation mode of the source-network-load-storage integrated system, screen out the operation mode with the largest matching degree as the scheduling plan, specifically including: Based on the real-time operation vector matrix, calculate the score recursion value of each operation mode per unit time; Verify the matching degree between the score recursion value and the interval vector range of the operation mode to obtain a matching degree value; Based on the matching degree value, screen out the operation mode with the largest matching degree as the scheduling plan; Among them, through the formula quantize the real-time operation vector matrix to obtain the real-time parameter value ; is the original value of the th parameter at time t; is the historical mean of the th parameter, which is used to center the data; is the historical standard deviation of the th parameter, which is used to scale the data distribution; Through the formula The score recursion value is calculated ; In the formula, the score recursion value represents the dynamic score of the th time unit for the th operation mode, reflecting the matching degree between real-time data and historical patterns. The lower the score, the higher the matching degree; and are the upper and lower limits of the interval vector range respectively, is the fitting value of real-time data, λ is the weight coefficient, 0 < λ < 1, reflecting the synergistic effect between the predicted value and the real-time parameter value; When verifying the matching degree between the score recursion value and the range of the interval vector of the running mode, if the real-time parameter falls within the range of the interval vector the matching degree value is 1; otherwise it is 0; Calculate real-time parameters Distance to the midpoint of the interval vector range , calculated through the formula ; The closer the value is to 1, the higher the matching degree; Combining the score recursion value and the interval matching result, the final matching degree value is calculated by weighting through the formula ; where is the interval matching weight ; in the formula Select The largest mode as the scheduling scheme.

2. The low-carbon optimization scheduling method based on source-network-load-storage integration according to claim 1, wherein, The specific method of calculating the correlation coefficient between each operation parameter and the operation mode of the source-network-load-storage integrated system based on the historical power demand, renewable energy generation, energy storage device status, and grid load data specifically includes: Based on the operation mode of the source-network-load-storage integrated system, mark the key operation parameters concerned by the system; Assign a weighted average value to each key operation parameter; Perform standardization processing on the historical power demand, renewable energy generation, energy storage device status, and grid load data to construct a scatter plot of historical operation data; Based on the scatter plot of historical operation data, key operation parameters, and weighted average value, calculate the correlation coefficient between each operation parameter and the key operation parameter in the historical operation data per unit time.

3. The low-carbon optimization scheduling method based on source-network-load-storage integration according to claim 2, wherein, The correlation coefficient is specifically as follows: ; in the formula, is the correlation coefficient between the th key operating parameter and the th operating parameter, is the value of the th operating parameter, is the weighted mean of the key operating parameters, is the total number of operating parameters.

4. A low-carbon optimal scheduling method based on source-network-load-storage integration according to claim 3, characterized in that Cluster the correlation coefficients between each operation parameter and the operation mode of the source-network-load-storage integrated system to obtain an array of correlation characteristic parameters of the system operation mode, specifically including: Based on the correlation coefficients between each operation parameter and the source-network-load-storage integrated system, establish a correlation coefficient matrix; Perform iterative clustering through the mean clustering algorithm: Step a Initialization: Randomly select K cluster centers from the correlation coefficient matrix as the initial cluster centroids ; i represents the index of the i-th cluster, i = 1, 2, …, K; Step b: Allocate data points: Calculate each data point and the current cluster centroid to obtain the Euclidean distance. Assign each point to the nearest cluster to form a new set of clusters ; k represents the number of iterations, that is, the clustering algorithm is currently in the k-th round of iteration; Step c Update cluster centers: Recalculate the centroid of each cluster ; is the number of data points in the i-th cluster; Step d iteration termination: Repeat steps b - c until the centroid change is less than the threshold value or the maximum number of iterations is reached. At this time, the centroid is the array of associated feature parameters of the system operation mode.

Images