Source network load storage flexibility quantification method based on simulation modeling analysis

By performing scenario sampling and adaptive weight calculation on the standardized sequence of source-grid-load-storage combination, the simulation modeling process was optimized, solving the problems of deviation in flexibility quantification results and low computational efficiency in existing technologies, and achieving higher accuracy and efficiency in flexibility assessment.

CN123116582APending Publication Date: 2026-11-17RES INST OF ECONOMICS & TECH STATE GRID SHANDONG ELECTRIC POWER
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611272340.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-21
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing simulation modeling schemes fail to effectively combine the actual operating characteristics of source-grid-load-storage equipment, resulting in large deviations in flexibility quantification results, low computational efficiency, and failure to adapt to the output fluctuation characteristics under different scenarios, thus failing to achieve accurate flexibility assessment.

Method used

By sampling the standardized sequence of source-grid-load-storage combination, initial scenario parameters are generated. Adjustment indicators are calculated based on power coupling relationship, adaptive weights are determined using information entropy, weighted summation is performed, and the simulation quantization process is optimized by combining regularization constant and iterative correction.

Benefits of technology

It improves the accuracy and efficiency of quantitative assessment of source-grid-load-storage flexibility, reduces computation time, and achieves more accurate flexibility assessment and higher simulation efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123116582A_ABST
    Figure CN123116582A_ABST
Patent Text Reader

Abstract

This invention relates to the field of power system technology and discloses a method for quantifying the flexibility of power generation, grid, load, and storage based on simulation modeling analysis. The method includes: firstly, extracting standardized sequence statistical characteristic parameters of power generation, grid, load, and storage to generate initial scenario parameters, and mapping power coupling relationships to obtain joint simulation data; then, calculating adjustable resource adjustment indicators based on the simulation data; using the information entropy of the indicators to obtain adaptive weights, and weighted summing to obtain the comprehensive quantification result of the system, completing a single quantification simulation; comparing the deviation between the quantification result and the preset expected value, and correcting the power coupling statistical characteristic parameters in reverse; repeating the simulation quantification process with the corrected parameters, iterating until the comprehensive quantification result converges, and finally outputting the converged quantification result to achieve quantitative evaluation of the flexibility of power generation, grid, load, and storage. This invention can improve the efficiency of quantifying the flexibility of power generation, grid, load, and storage based on simulation modeling analysis.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of power system technology, and in particular to a method for quantifying the flexibility of power generation, grid, load and storage based on simulation modeling and analysis. Background Technology

[0002] With the advancement of new power system construction, the deep coupling of diverse adjustable resources such as power generation, grid, load, and storage has made simulation modeling-based quantitative assessment of system flexibility a core technical means for power grid dispatching and planning. Existing simulation quantification schemes mostly adopt a fixed-weight index weighting system, failing to dynamically allocate weights based on the dispersion of each adjustment index's numerical distribution. A single fixed weight cannot adapt to the power output fluctuation characteristics of power generation, grid, load, and storage under different time-series scenarios, leading to significant deviations in the comprehensive flexibility quantification results. Furthermore, traditional simulation processes employ a single open-loop calculation mode, lacking a feedback correction mechanism for deviations between quantification results and preset expectations. The statistical characteristic parameters of power coupling relationships cannot adaptively adjust with quantification errors, resulting in poor convergence of simulation calculations, numerous repeated simulation iterations, and a long overall quantification calculation time.

[0003] Existing technologies, when constructing joint simulation datasets for power generation, grid, load, and storage, simply overlay the power time-series sequences of various devices without considering the grid topology to differentiate the power transmission, charging, discharging, and feedback path constraints between source, load, and storage nodes. They also ignore the constraints of line capacity and transmission losses on the actual adjustment capabilities of adjustable resources. Furthermore, when sampling to generate scenario parameters, they do not construct stratified sampling intervals based on time-series mean, variance, and kurtosis, resulting in low matching between scenario samples and actual grid operating conditions. In addition, existing flexibility capacity calculation models do not introduce regularization constants to suppress computational distortion caused by minimal power differences, leading to inaccurate calculations of the adjustment potential of adjustable resources at different time points. This further reduces the accuracy of quantitative assessment of power generation, grid, load, and storage flexibility and the efficiency of simulation operations. Therefore, how to balance the accuracy of quantitative assessment of power generation, grid, load, and storage flexibility with the efficiency of simulation operations has become an urgent technical problem to be solved in the field of new power system dispatching and planning. Summary of the Invention

[0004] This invention provides a source-grid-load-storage flexibility quantification method based on simulation modeling analysis to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides a source-grid-load-storage flexibility quantification method based on simulation modeling analysis, comprising: Ⅰ. Scenario sampling is performed on the statistical characteristic parameters of the standardized sequence in the source-grid-load-storage combination to generate the initial scenario parameters of the standardized sequence, and the initial scenario parameters are mapped to a preset power coupling relationship to obtain the joint simulation data corresponding to the initial scenario parameters; II. Calculate the adjustment index of the adjustable resources corresponding to the joint simulation data; III. Determine the adaptive weight of the corresponding index by the information entropy of the adjustment index value, and sum the adjustment index values ​​according to the adaptive weight to obtain the comprehensive quantization result of the source-grid-load-storage combination, thus completing one simulation quantization process; IV. Based on the deviation between the comprehensive quantization result and the preset expected value, the statistical characteristic parameters in the power coupling relationship are reverse-corrected to obtain the corrected statistical characteristic parameters; V. Based on the corrected statistical characteristic parameters, iterate the simulation quantization process until the comprehensive quantization result converges, and output the comprehensive quantization result at the convergence point.

[0006] In a preferred embodiment, before performing scenario sampling on the statistical characteristic parameters of the standardized sequence in the source-grid-load-storage combination, the following steps are included: The ratio of the actual operating power value to the corresponding rated power value of the adjustable equipment in the source-grid-load-storage combination is used as the standardized power value of the adjustable equipment at each time section, and the standardized power values ​​are arranged in time sequence to obtain the standardized power sequence corresponding to the adjustable equipment. The standardized power sequences are grouped and merged according to equipment type to obtain the standardized equipment sequence of the source-grid-load-storage combination. Sliding window truncation is performed on the standardized sequence of the device, and the arithmetic mean, variance and kurtosis of the standardized sequence in each sliding window are used as the statistical feature sub-parameters of the sliding window; The statistical feature sub-parameters corresponding to the same equipment type are concatenated according to the window time sequence to obtain the statistical feature parameters corresponding to each type of equipment.

[0007] In a preferred embodiment, the step of sampling the statistical characteristic parameters of the standardized sequence in the source-grid-load-storage combination for scenarios to generate initial scenario parameters of the standardized sequence, and mapping the initial scenario parameters to a preset power coupling relationship to obtain the joint simulation data corresponding to the initial scenario parameters, includes: The mean sequence, variance sequence, and kurtosis sequence of the equipment type are read from the statistical feature parameters. The values ​​at each time point in the mean sequence are used as the midline of the interval, the values ​​at each time point in the variance sequence are used as the basic radius length, and the values ​​at each time point in the kurtosis sequence are used as the boundary correction coefficients to construct the sampling interval of the equipment type. Within the sampling interval, stratified equidistant sampling is performed on the device type to obtain the initial scene sub-parameters of the device type at each time point; The initial scene sub-parameters are merged according to device type identifier and time identifier to obtain the initial scene parameters of the standardized sequence; Based on the initial scenario parameters, the reference power mapping value of each device type at each time is found from the preset power coupling relationship, and the power offset corresponding to the sampled value in the initial scenario parameters is superimposed on the reference power mapping value to obtain the power mapping value of the device type. By merging the power mapping values ​​of all device types at each time point, the joint simulation data corresponding to the initial scenario parameters is obtained.

[0008] In a preferred embodiment, the step of finding the reference power mapping value for each device type at each time from a preset power coupling relationship includes: Adjustable devices with values ​​of source-side, load-side, and storage-side respectively are selected from the device type field of the adjustable devices and designated as source-side devices, load-side devices, and storage devices, respectively. The access nodes of the source-side device, load-side device, and storage-side device are respectively marked as source node, load node, and storage node, and the node number of each associated storage node is bound with the corresponding device type identifier to generate the node device association index of the adjustable device; Using the source node, the load node, and the storage node as vertices of the graph, and the transmission lines and transformers between the nodes as edges of the graph, an undirected topology graph reflecting the node connection relationships is constructed. First, using the source node as the starting vertex and the load node or the storage node as the target vertex, then using the storage node as the starting vertex and the load node as the target vertex, perform a breadth-first traversal on the undirected topology graph to search for all connected paths between the starting vertex and the target vertex.

[0009] In a preferred embodiment, the step of finding the reference power mapping value for each device type at each time from a preset power coupling relationship includes: The connected path with the source node as the starting vertex and the load node as the target vertex is recorded as the power transmission path; the connected path with the storage node as the target vertex is recorded as the power charging and discharging path; and the connected path with the storage node as the starting vertex is recorded as the power feedback path. Assign a path identifier to each connected path found, and bind the path identifier to the nodes at both ends of the power transmission path, the power charging and discharging path, and the power feedback path to generate a path node association index for each path; Based on the rated voltage level of each node, the maximum transmission capacity and transmission loss rate are configured for each path, and the path parameter record set is written according to the path identifier of each path to generate the path parameter index of each path. The statistical feature parameters are associated with the node device association index, the path node association index, and the path parameter index, and compiled to generate a power reference mapping value for each device type identifier. A power coupling relationship is constructed using the device type identifier, node number, path identifier, and time segment as the joint primary key, and the power reference mapping value as the attribute value.

[0010] In a preferred embodiment, calculating the adjustment index of the adjustable resources corresponding to the co-simulation data includes: Read the actual power values ​​of the adjustable device at each time section from the joint simulation data, and read the upper limit and lower limit of the rated power of the adjustable device. The difference between the actual power value and the corresponding upper limit of rated power and lower limit of rated power is obtained by subtracting them from the actual power value, respectively, to obtain the power adjustment difference and the power adjustment difference at each time section. The comprehensive flexibility capacity value of the adjustable equipment at each time segment is calculated based on the upward power adjustment difference and the downward power adjustment difference. The formula for calculating the comprehensive flexibility capacity value is as follows: ; in, Indicates the first The adjustable device in the first The comprehensive flexibility capacity value under a specific time segment. Indicates the first The adjustable device in the first The difference in upward power across different time segments Indicates the first The adjustable device in the first The difference in downward power across different time segments This represents the preset regularization constant; The comprehensive flexibility capacity values ​​are arranged in chronological order to generate a comprehensive flexibility capacity sequence for each adjustable device, and the comprehensive flexibility capacity sequence is used as an adjustment index for the adjustable resources corresponding to the joint simulation data.

[0011] In a preferred embodiment, the step of determining the adaptive weight of the corresponding index through the information entropy of the adjustment index value, and then weighting and summing the adjustment index values ​​according to the adaptive weight to obtain the comprehensive quantization result of the source-grid-load-storage combination, thus completing one simulation quantization process, includes: The numerical sequence is divided into several equally wide numerical intervals according to the numerical distribution range in the numerical sequence corresponding to the adjustment index, and the number of values ​​falling into each numerical interval is counted to obtain the frequency distribution vector corresponding to each adjustment index. Each frequency in the frequency distribution vector is compared with the total frequency of the corresponding adjustment index to obtain the distribution probability vector of the adjustment index, and the ratio of the number of non-zero components in the distribution probability vector to the total number of components is used as the discreteness characterization value of the adjustment index. The dispersion characterization value is used as the weight base value of the adjustment index, and the ratio of the weight base value of the adjustment index to the sum of the weight base values ​​of all adjustment indices is used as the adaptive weight of the adjustment index. The product of the adjustment index and the corresponding adaptive weight is used as the weighted component of each adjustment index at each time section, and the sum of the weighted components is used as the comprehensive quantitative value of each time section. The comprehensive quantization values ​​of the cross-sections under all time sections are arranged in chronological order to obtain the comprehensive quantization result of the source-grid-load-storage combination, thus completing one simulation quantization process.

[0012] In a preferred embodiment, the step of determining the deviation between the comprehensive quantification result and the preset expected value includes: The initial deviation value of the source-grid-load-storage combination under each time section is obtained by subtracting the comprehensive quantitative value of each time section from the expected threshold of the corresponding time section in the preset expected value. The initial deviation values ​​are classified and labeled according to their positive or negative signs. The initial deviation values ​​with positive values ​​are labeled as positive deviation values, and the initial deviation values ​​with negative values ​​are labeled as negative deviation values. The positive and negative deviation values ​​at each time section are arranged in chronological order to obtain a deviation value sequence.

[0013] In a preferred embodiment, the step of reverse-correcting the statistical characteristic parameters in the power coupling relationship to obtain the corrected statistical characteristic parameters includes: A positive deviation window is determined based on the continuous distribution interval of the positive deviation value over time, and a negative deviation window is determined based on the continuous distribution interval of the negative deviation value over time. The statistical characteristic parameters corresponding to the positive deviation window are read from the power coupling relationship as the set of parameters to be reduced, and the statistical characteristic parameters corresponding to each time section in the negative deviation window are read from the power coupling relationship as the set of parameters to be increased. The statistical characteristic parameters in the set of parameters to be reduced are compressed along the direction of the preset lower threshold to obtain the set of down-adjustment correction parameters for the positive deviation window; The statistical characteristic parameters in the set of parameters to be increased are stretched along the preset upper threshold direction to obtain the set of adjustment correction parameters for the negative deviation window; The down-modulation correction parameter set and the up-modulation correction parameter set are re-filled into the power coupling relationship according to their respective time sections to obtain the corrected statistical characteristic parameters in the power coupling relationship.

[0014] In a preferred embodiment, the step of iterating the simulation quantization process based on the corrected statistical characteristic parameters until the comprehensive quantization result converges, and then outputting the comprehensive quantization result at the point of convergence, includes: The corrected statistical characteristic parameters are substituted into the simulation quantization process for iteration to obtain the updated comprehensive quantization result. The updated comprehensive quantization result is subtracted from the comprehensive quantization result obtained in the previous iteration at the same time section, and the absolute value is taken to obtain the fluctuation amplitude of the source-grid-load-storage combination at each time section. When the fluctuation amplitude at any time segment exceeds the preset convergence tolerance, the updated comprehensive quantization result is used as the comparison benchmark for the next iteration, and the reverse correction operation is re-executed until the fluctuation amplitude at all time segments meets the condition of being less than or equal to the convergence tolerance. When the fluctuation amplitude under all time sections is less than or equal to the convergence tolerance, it is determined that the updated comprehensive quantization result has reached the convergence state, and the updated comprehensive quantization result is marked as the comprehensive quantization result at the time of convergence. The convergence results are arranged according to time sections to obtain the final convergence result, which is then output.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. This scheme relies on a time-series sliding window to extract multi-dimensional statistical features of the standardized power sequence, including mean, variance, and kurtosis. It combines these mean, variance, and kurtosis parameters to construct stratified sampling intervals, generating initial scenario parameters that closely match actual operating conditions. Based on the power grid topology, it constructs power coupling relationships that distinguish between transmission, charging / discharging, and feedback types, fully considering line transmission capacity and loss constraints to generate joint simulation data. A comprehensive flexibility capacity calculation model with regularization constants is introduced to eliminate calculation distortion caused by minimal power differences. The scheme uses the information entropy of adjustment indicators to quantify the numerical dispersion and adaptively allocate indicator weights. Weighted summation of time-series sections yields refined comprehensive quantitative results, effectively improving the calculation accuracy of the source-grid-load-storage flexibility quantitative assessment.

[0016] 2. This scheme establishes a closed-loop iterative correction architecture for quantization results. It achieves directional adjustment of power coupling statistical characteristic parameters through time-series deviation window division. The iteration termination condition is determined by the cross-sectional fluctuation amplitude and convergence tolerance. The parameter adaptive correction mechanism reduces invalid simulation loops and shortens the number of iterations required for quantization result convergence. The entire simulation quantization process realizes fully automated calculation of scenario generation, index calculation, weight assignment, parameter correction, and convergence determination, significantly improving the overall computational efficiency of source-grid-load-storage flexibility simulation quantization. Attached Figure Description

[0017] Figure 1 A flowchart illustrating a source-grid-load-storage flexibility quantification method based on simulation modeling analysis provided in an embodiment of the present invention; Figure 2 This is a timing simulation result of the system feature evaluation stage provided in an embodiment of the present invention; Figure 3 This is a time series curve of source-load-storage flexibility index stabilization provided in an embodiment of the present invention for the feature extraction stage; Figure 4 This is a timing curve for tracking the index of load-side negative deviation-triggered compression calibration provided in an embodiment of the present invention; Figure 5 This is a time-series curve of the entire convergence process of the flexibility index under multi-round iterative calibration provided in an embodiment of the present invention; Figure 6 The simulation curves for the multi-round deviation calibration iteration process provided in an embodiment of the present invention are shown.

[0018] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0019] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0020] This application provides a method for quantifying source-grid-load-storage flexibility based on simulation modeling analysis. The execution entity of this method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the method can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cluster of cloud servers. The server can be an independent server or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0021] Reference Figure 1 The diagram shown is a flowchart illustrating a source-grid-load-storage flexibility quantification method based on simulation modeling analysis according to an embodiment of the present invention. In this embodiment, the source-grid-load-storage flexibility quantification method based on simulation modeling analysis includes: Ⅰ. Scenario sampling is performed on the statistical characteristic parameters of the standardized sequence in the source-grid-load-storage combination to generate the initial scenario parameters of the standardized sequence, and the initial scenario parameters are mapped to a preset power coupling relationship to obtain the joint simulation data corresponding to the initial scenario parameters; In this embodiment of the invention, before performing scenario sampling on the statistical characteristic parameters of the standardized sequence in the source-grid-load-storage combination, the following steps are included: The ratio of the actual operating power value to the corresponding rated power value of the adjustable equipment in the source-grid-load-storage combination is used as the standardized power value of the adjustable equipment at each time section, and the standardized power values ​​are arranged in time sequence to obtain the standardized power sequence corresponding to the adjustable equipment. The standardized power sequences are grouped and merged according to equipment type to obtain the standardized equipment sequence of the source-grid-load-storage combination. Sliding window truncation is performed on the standardized sequence of the device, and the arithmetic mean, variance and kurtosis of the standardized sequence in each sliding window are used as the statistical feature sub-parameters of the sliding window; The statistical feature sub-parameters corresponding to the same equipment type are concatenated according to the window time sequence to obtain the statistical feature parameters corresponding to each type of equipment.

[0022] The step of sampling statistical characteristic parameters of the standardized sequence in the source-grid-load-storage combination for scenarios to generate initial scenario parameters of the standardized sequence, and mapping the initial scenario parameters to a preset power coupling relationship to obtain joint simulation data corresponding to the initial scenario parameters, includes: The mean sequence, variance sequence, and kurtosis sequence of the equipment type are read from the statistical feature parameters. The values ​​at each time point in the mean sequence are used as the midline of the interval, the values ​​at each time point in the variance sequence are used as the basic radius length, and the values ​​at each time point in the kurtosis sequence are used as the boundary correction coefficients to construct the sampling interval of the equipment type. Within the sampling interval, stratified equidistant sampling is performed on the device type to obtain the initial scene sub-parameters of the device type at each time point; The initial scene sub-parameters are merged according to device type identifier and time identifier to obtain the initial scene parameters of the standardized sequence; Based on the initial scenario parameters, the reference power mapping value of each device type at each time is found from the preset power coupling relationship, and the power offset corresponding to the sampled value in the initial scenario parameters is superimposed on the reference power mapping value to obtain the power mapping value of the device type. By merging the power mapping values ​​of all device types at each time point, the joint simulation data corresponding to the initial scenario parameters is obtained.

[0023] The step of finding the reference power mapping value for each device type at each time from the preset power coupling relationship includes: Adjustable devices with values ​​of source-side, load-side, and storage-side respectively are selected from the device type field of the adjustable devices and designated as source-side devices, load-side devices, and storage devices, respectively. The access nodes of the source-side device, load-side device, and storage-side device are respectively marked as source node, load node, and storage node, and the node number of each associated storage node is bound with the corresponding device type identifier to generate the node device association index of the adjustable device; Using the source node, the load node, and the storage node as vertices of the graph, and the transmission lines and transformers between the nodes as edges of the graph, an undirected topology graph reflecting the node connection relationships is constructed. First, using the source node as the starting vertex and the load node or the storage node as the target vertex, then using the storage node as the starting vertex and the load node as the target vertex, perform a breadth-first traversal on the undirected topology graph to search for all connected paths between the starting vertex and the target vertex.

[0024] The step of finding the reference power mapping value for each device type at each time from the preset power coupling relationship also includes: The connected path with the source node as the starting vertex and the load node as the target vertex is recorded as the power transmission path; the connected path with the storage node as the target vertex is recorded as the power charging and discharging path; and the connected path with the storage node as the starting vertex is recorded as the power feedback path. Assign a path identifier to each connected path found, and bind the path identifier to the nodes at both ends of the power transmission path, the power charging and discharging path, and the power feedback path to generate a path node association index for each path; Based on the rated voltage level of each node, the maximum transmission capacity and transmission loss rate are configured for each path, and the path parameter record set is written according to the path identifier of each path to generate the path parameter index of each path. The statistical feature parameters are associated with the node device association index, the path node association index, and the path parameter index, and compiled to generate a power reference mapping value for each device type identifier. A power coupling relationship is constructed using the device type identifier, node number, path identifier, and time segment as the joint primary key, and the power reference mapping value as the attribute value.

[0025] All adjustable devices within the source-grid-load-storage combination are identified, and all fixed time segments corresponding to the entire set of time-series operation data are determined. The actual operating power value generated by each adjustable device under each fixed time segment is collected. The rated power value fixed in the equipment parameter file of each adjustable device is retrieved. Taking a single time segment and a single device as the calculation unit, the proportion of the actual operating power of the device under the current time segment is calculated using the device's predetermined rated power value. This proportion is the standardized power value of the adjustable device under a single time segment. After collecting the standardized power values ​​of all fixed time segments of a single adjustable device, all standardized power values ​​are arranged in order according to the fixed sorting rules of the time segment sequence. The complete time-series value set formed after sorting is the standardized power sequence corresponding to the adjustable device.

[0026] Based on the equipment category classification standards preset in the source-grid-load-storage system, the standardized power sequences generated by all independent adjustable equipment are classified and integrated. All standardized power sequences corresponding to equipment with the same category attributes are grouped into the same data group, completing the classification and merging operation of all standardized power sequences. The overall set of standardized power sequences after all classification and merging is uniformly defined as the standardized equipment sequence of the source-grid-load-storage combination.

[0027] For each independent sequence in the merged standardized equipment sequence, a sliding window truncation method with a fixed window length and fixed single-step sliding distance is used to segment the sequence data. The single sliding distance of the sliding window is fixed at one standard time segment unit to ensure that the window slides continuously position by position on the time sequence without jumps, overlaps, or omissions. After each sliding pause, all standardized sequence data within the current window's coverage area are locked, and a global numerical summation operation is performed on all data within the window. The total summation value is then divided by the total number of standardized sequence data contained in the current window to obtain the result. The arithmetic mean of the standardized sequence within the current sliding window is calculated by squaring the difference between each standardized sequence value and its corresponding arithmetic mean. The average of all squaring results is then calculated, resulting in the variance of the standardized sequence within the current sliding window. Based on the distribution pattern of the standardized sequence values ​​within the window, the kurtosis of the distribution is calculated, resulting in the kurtosis of the standardized sequence within the current sliding window. The overall set of arithmetic mean, variance, and kurtosis values ​​calculated for a single sliding window constitutes the statistical characteristic sub-parameter of that sliding window.

[0028] Using the classification results of equipment type as a fixed grouping basis, the statistical feature sub-parameters generated by each sliding window under the same equipment type are sequentially spliced ​​and integrated according to the corresponding temporal order of the sliding window during the sequence truncation process. The statistical feature sub-parameters of the windows with earlier temporal order are prioritized and the statistical feature sub-parameters of the windows with later temporal order are sequentially arranged. After completing the temporal splicing and integration of all statistical feature sub-parameters of the same equipment type, the complete time-series feature data set formed is the statistical feature parameter corresponding to each type of equipment. This set of parameters provides preliminary basic data for the scenario sampling work of standardized sequences in the source-grid-load-storage combination.

[0029] The statistical characteristic parameters corresponding to various types of equipment are retrieved. From the complete statistical characteristic parameter dataset, the mean sequence, variance sequence, and kurtosis sequence specific to each equipment type are extracted separately. Each time-series value within the mean sequence is used as the baseline positioning value for the interval center. Each time-series value within the variance sequence is used as the basic scale value for the interval range. Each time-series value within the kurtosis sequence is used as the fine-tuning limit value for the interval boundary range. By combining the time-series values ​​of the three types of sequences, the sampling interval covering the full time-series data of the corresponding equipment type is constructed. The constructed interval structure is the sampling interval for the corresponding equipment type.

[0030] Using pre-built sampling intervals for different device types with independent time series as a fixed sampling range, stratified equidistant sampling is performed for each time series sampling interval corresponding to a single device type. The stratification is based on equal-level intervals divided according to the distribution of time series data within the sampling interval. The equidistant sampling uses a fixed and equal point interval within each level interval as the sampling benchmark. A fixed number of valid sample points are extracted within each level interval. All corresponding sample data obtained after sampling is completed under all time series constitute the initial scene sub-parameters of the device type under each time segment.

[0031] The initial scene sub-parameters generated by all device types at each time point are collected and organized. The unique device type identifier preset by the system is used as the classification and collection basis, and the fixed time identifier is used as the time sequence sorting basis. The initial scene sub-parameters belonging to the same device type but at different times are organized and collected in time sequence. The organized data of different device types are integrated into a complete dataset. The complete parameter set obtained after integration and organization is the initial scene parameter of the standardized sequence.

[0032] Using the fully generated initial scenario parameters as the sole retrieval basis, the system retrieves the power coupling relationship dataset, which is pre-configured and stored to adapt to the operating rules of source-grid-load-storage equipment. It then matches and locates the baseline power mapping value corresponding to each time node for each equipment type. This baseline power mapping value is a fixed mapping reference value under the standard operating state of the equipment. The sample values ​​of each sampling point in the initial scenario parameters correspond to a unique power offset. This power offset is a fixed offset value obtained by matching the deviation between the sampled values ​​and the standard statistical values. The corresponding power offset is superimposed on the matched baseline power mapping value. After completing the numerical correction, the result is the power mapping value for each time point of the corresponding equipment type.

[0033] By traversing all time points, the power mapping values ​​generated for all different device types at a single time point are uniformly merged and collected. After completing the power mapping value merging operation for all time points, the resulting integrated data set that is fully covered, has complete timing, and includes all device types is the joint simulation data corresponding to the initial scenario parameters.

[0034] The system retrieves complete archive data of all adjustable devices that have been filed and stored in the system. It reads the preset fixed format device type field value in the data entry of each adjustable device one by one. Using the predefined fixed source-side standard field value, load-side standard field value, and storage-side standard field value as the unique screening and matching benchmark, all adjustable devices whose device type field value is completely consistent with the source-side standard field value are uniformly grouped into source-side devices, all adjustable devices whose device type field value is completely consistent with the load-side standard field value are uniformly grouped into load-side devices, and all adjustable devices whose device type field value is completely consistent with the storage-side standard field value are uniformly grouped into storage-side devices. This achieves accurate classification of all adjustable devices according to the power grid functional attributes.

[0035] All marked source nodes, load nodes, and storage nodes are uniformly set as independent vertex units of the topology. Transmission lines that carry out power transmission between nodes and transformers that carry out voltage level transformation and power transfer within the power system are uniformly set as connection edge units of the topology. According to the fixed connection relationship of the actual physical wiring of the power grid, the precise docking and matching of all vertex units and connection edge units is completed. The constructed topology eliminates the unidirectional constraint of power transmission direction and can objectively restore the real physical connection structure between each power node. This topology is defined as an undirected topology graph.

[0036] The grid connection points corresponding to the selected source-side equipment, load-side equipment, and storage-side equipment are fixedly marked. The grid connection nodes of the source-side equipment are uniformly marked as source nodes, the grid connection nodes of the load-side equipment are uniformly marked as load nodes, and the grid connection nodes of the storage-side equipment are uniformly marked as storage nodes. The unique fixed node number of each connection node and the unique equipment type identifier of the corresponding equipment are retrieved, and the two sets of data are linked and archived in a one-to-one binding association storage method to form a node equipment association index of adjustable equipment that can be independently retrieved and accurately matched.

[0037] Using the completed undirected topology graph as a fixed retrieval carrier, all source nodes contained in the undirected topology graph are selected as the starting vertices for traversal, and all load nodes and storage nodes contained in the undirected topology graph are selected as the target vertices for traversal. Based on the fixed rule of breadth-first traversal and layer-by-layer neighbor retrieval, the retrieval range is expanded outward layer by layer from the directly adjacent nodes of the starting vertex, and all connected paths with complete physical connectivity between the starting vertex and the target vertex are investigated and counted.

[0038] Continuing with the completed undirected topology graph as the fixed retrieval carrier, we select all storage nodes contained in the undirected topology graph as the starting vertex for traversal and all load nodes contained in the undirected topology graph as the target vertex for traversal. We continue to expand the retrieval range layer by layer based on the fixed rule of breadth-first traversal and layer-by-layer neighbor retrieval. We investigate and count all connected paths with complete physical connectivity between the current starting vertex and the target vertex. Finally, we summarize the set of all valid connected paths in the undirected topology graph that meet the connectivity conditions.

[0039] All connected paths obtained after traversing the undirected topology graph are classified, named, and archived. All connected paths with the starting vertex as the source node and the target vertex as the load node are uniformly classified and recorded as power transmission paths. All connected paths with the starting vertex as the source node and the target vertex as the storage node are uniformly classified and recorded as power charging and discharging paths. All connected paths with the starting vertex as the storage node and the target vertex as the load node are uniformly classified and recorded as power feedback paths. All path classifications are strictly defined based on the fixed node attributes of the two ends of the path to ensure that the classification results of each type of path are unique and distinguishable.

[0040] For each connected path within the topology that has been classified and defined, a unique and non-repeating path identifier is assigned under the system's preset rules. This path identifier serves as an independent identification code for a single path. Each path identifier is then bound and aggregated with the topology node information at both ends of the corresponding power transmission path, power charging / discharging path, and power feedback path. This completes the one-to-one correspondence and integration of all paths with their corresponding endpoint nodes. The dataset after all associations and integrations is then uniformly summarized to generate a path node association index that covers all connected path information.

[0041] The rated voltage level parameters stored in the equipment files of the access nodes at both ends of the path are retrieved. The fixed rated voltage level of the node is used as the sole matching benchmark for the path transmission performance configuration. Based on the power grid equipment operation standards corresponding to different rated voltage levels, a fixed maximum transmission capacity and a fixed transmission loss rate are matched for each independent connected path. The configured maximum transmission capacity and transmission loss rate are entered into the path parameter record set one by one according to the unique path identifier of each path. After all path parameters are entered and integrated, a path parameter index for each path that can be independently searched and corresponds one-to-one is formed.

[0042] The statistical feature parameters of various devices generated in the previous process are retrieved. The device time-series feature data of the statistical feature parameters are then combined with the completed node device association index, path node association index, and path parameter index to carry out comprehensive data association and compilation processing. Based on the data association relationship, the device affiliation relationship, the node access relationship, and the path transmission parameter relationship, the feature data, node data, and path data are accurately matched and integrated. After the data compilation and integration is completed, the power reference mapping value corresponding to each independent device type identifier is generated.

[0043] The four data items—device type identifier, node number, path identifier, and time section—are collectively set as the joint primary key for data retrieval and storage. The four primary key data items are combined in pairs and mutually constrained to ensure that each data entry has unique identifiability. The power reference mapping value generated by the corresponding device type identifier is set as the attribute value corresponding to the joint primary key. The dataset is built based on the corresponding storage structure of fixed primary key and attribute value, and finally forms a power coupling relationship that can comprehensively cover the dimensions of device, node, path, and time series.

[0044] The beneficial effects include: by finely classifying and defining power grid connection paths, distinguishing between power transmission paths, power charging and discharging paths, and power feedback paths, the power interaction logic of source, load, and storage devices can be accurately differentiated, eliminating data confusion in different power transmission scenarios. By configuring unique path identifiers for each path and constructing path node association indexes and path parameter indexes, precise binding and correspondence between devices, nodes, and paths are achieved, solving the technical defects of traditional power grid simulation data being messy and having low matching accuracy. Based on matching the maximum transmission capacity and transmission loss rate corresponding to the rated voltage level of nodes, parameters can be configured to strictly conform to the actual operating conditions of the power grid, ensuring that parameter configurations conform to real physical characteristics and improving data authenticity and engineering adaptability. By generating power benchmark mapping values ​​through the association and compilation of multi-dimensional indexes and statistical feature parameters, and using multi-dimensional joint primary keys to construct power coupling relationships, unified coupled modeling of time series, devices, nodes, and paths is achieved, effectively improving the overall integrity and accuracy of power scenario simulation of source-grid-load-storage systems, and ensuring that subsequent scenario sampling and simulation data analysis results have high reproducibility and engineering applicability.

[0045] II. Calculate the adjustment index of the adjustable resources corresponding to the joint simulation data; In this embodiment of the invention, calculating the adjustment index of the adjustable resource corresponding to the joint simulation data includes: Read the actual power values ​​of the adjustable device at each time section from the joint simulation data, and read the upper limit and lower limit of the rated power of the adjustable device. The difference between the actual power value and the corresponding upper limit of rated power and lower limit of rated power is obtained by subtracting them from the actual power value, respectively, to obtain the power adjustment difference and the power adjustment difference at each time section. The comprehensive flexibility capacity value of the adjustable equipment at each time segment is calculated based on the upward power adjustment difference and the downward power adjustment difference. The formula for calculating the comprehensive flexibility capacity value is as follows: ; in, Indicates the first The adjustable device in the first The comprehensive flexibility capacity value under a specific time segment. Indicates the first The adjustable device in the first The difference in upward power across different time segments Indicates the first The adjustable device in the first The difference in downward power across different time segments This represents the preset regularization constant; The comprehensive flexibility capacity values ​​are arranged in chronological order to generate a comprehensive flexibility capacity sequence for each adjustable device, and the comprehensive flexibility capacity sequence is used as an adjustment index for the adjustable resources corresponding to the joint simulation data.

[0046] Retrieve the fully generated co-simulation data set, extract the actual power values ​​of each adjustable device at all preset fixed time sections from the co-simulation data, and simultaneously retrieve the pre-fixed upper and lower limits of rated power stored in the basic equipment file of each adjustable device, thus completing the complete extraction and collection of the three types of power values ​​of a single time node of a single adjustable device.

[0047] Using a single fixed time segment of an adjustable device as an independent accounting unit, the upper limit of the rated power corresponding to the same time segment is subtracted from the actual power value of the adjustable device under that time segment. After completing the difference accounting operation, the upward power difference corresponding to that time segment is obtained. The lower limit of the rated power is subtracted from the actual power value of the adjustable device corresponding to the same time segment. After completing the difference accounting operation, the downward power difference corresponding to that time segment is obtained. The upward and downward power differences matched by each time segment of a single adjustable device are obtained in a complete manner.

[0048] Using the power adjustment difference and power adjustment difference calculated at the same time segment as a unified accounting basis, the two sets of difference values ​​are integrated and summarized to form a complete capacity accounting basis. The adjustable power margin range corresponding to the two sets of difference values ​​is fully integrated and statistically analyzed. After the integration and statistical analysis is completed, the comprehensive flexibility capacity value of the adjustable equipment at a single time segment is obtained. The comprehensive flexibility capacity value calculation of a single adjustable equipment at all time segments is completed in sequence.

[0049] The comprehensive flexibility capacity values ​​calculated from all time segments of a single adjustable device are arranged in an orderly manner according to the chronological order of the time segments. After arrangement and integration, a comprehensive flexibility capacity sequence specific to each adjustable device is formed. The comprehensive flexibility capacity sequence generated by each adjustable device is uniformly defined as the adjustment index of the adjustable resources corresponding to the joint simulation data.

[0050] The comprehensive flexibility capacity value is the output data obtained after a single adjustable device completes all capacity calculation operations in a single time segment. This data is an intermediate product in the process of solving the adjustable resource adjustment index based on co-simulation data. All the prerequisite data required for the calculation operation comes from the upward and downward power difference obtained after splitting the co-simulation data.

[0051] The power difference adjustment is generated by calculating the difference based on the actual power values ​​of the adjustable equipment at each time segment extracted from the joint simulation data and the rated power upper limit value permanently archived in the equipment file. The difference calculation sets a single time segment of a single device as an independent calculation unit, and uses the rated power upper limit value of the corresponding time segment to subtract the actual power value of the adjustable equipment at the same time segment. After the calculation is completed, the complete value corresponding to the parameter is directly generated.

[0052] The power reduction difference is calculated by extracting the actual power values ​​of the adjustable equipment at each time segment from the joint simulation data and the rated power lower limit value permanently archived in the equipment file. The difference calculation sets a single time segment of a single device as an independent calculation unit, and uses the actual power value of the adjustable equipment at the same time segment to subtract the rated power lower limit value of the corresponding time segment. After the calculation is completed, the complete value corresponding to the parameter is directly generated.

[0053] The regularization constant is a constant that is pre-entered into the system and whose value is permanently fixed. The fixed value of this constant is determined based on the statistical conclusions of the historical simulation data accumulated by the source-grid-load-storage system over many years. All operating condition samples in the historical simulation scenarios where the power difference is adjusted upward and downward are simultaneously zero are selected as the sole criterion for determining the value. This constant is set to prevent the interruption of the data calculation process caused by all values ​​of the denominator being zero.

[0054] This numerical calculation method performs integrated quantitative statistical operations on the upward and downward power adjustment space of adjustable equipment under the same time segment. The value obtained by multiplying the two-way adjustment margins reflects the overall potential of the two-way power adjustment of the adjustable equipment. The denominator is superimposed with two types of adjustment margins and regularization constants to complete the normalization constraint processing of the quantitative results. The final comprehensive flexibility capacity value can simultaneously reflect the comprehensive performance of the upward and downward two-way power adjustment of a single adjustable equipment under the same time segment. The comprehensive flexibility capacity sequence formed by arranging the values ​​in chronological order can directly serve as the adjustment index of the adjustable resources corresponding to the joint simulation data. It fully matches the complete generation logic of the entire set of adjustable resource adjustment indexes, and provides a unified standard and quantitative judgment basis for the performance evaluation of various adjustable resources within the source-grid-load-storage system, supporting horizontal comparison.

[0055] The core of the formula for calculating the comprehensive flexibility capacity value is a weighted harmonic average model with a regularization term. From a physical perspective, this model accurately describes the bidirectional adjustment constraints of adjustable resources. When the difference in power adjustment or the difference in power adjustment in either direction approaches zero, the comprehensive flexibility capacity value also approaches zero, objectively reflecting the fundamental nature of power grid operation where the overall flexibility is lost due to the exhaustion of unidirectional adjustment margin.

[0056] From the perspective of dimensional consistency, the difference between upward and downward power adjustments both belong to the power dimension. The product of the two and the sum of the two results in a power unit, and there is no dimensional conflict.

[0057] The key to this model lies in the introduction of a regularization constant. This constant fundamentally eliminates the division-by-zero calculation interruption that occurs when both the upward and downward power difference are zero, ensuring that high-concurrency timing simulations can be executed stably under all operating conditions. On the other hand, when the power difference is in a very small range, this constant can effectively suppress non-physical distortions caused by numerical fluctuations, acting as a smoothing filter and significantly improving the continuity and reliability of the full-time-series adjustment index sequence.

[0058] Within the entire closed-loop iterative algorithm framework, the calculation formula for the comprehensive flexibility capacity value directly utilizes the joint simulation data generated by the previous power coupling mapping. The actual power values ​​at each time point are subtracted from the rated upper and lower limits to obtain the upward and downward adjustment differences. These are then input into the model to output a comprehensive flexibility capacity sequence with clear physical boundaries and continuous differentiability. This sequence provides fundamental data with a clear numerical distribution and quantifiable dispersion for subsequent adaptive weight allocation based on information entropy. Simultaneously, the capacity value output by this model can sensitively respond to changes in statistical characteristic parameters due to the reverse correction of deviations in the previous iteration. When the mean, variance, or kurtosis parameters in the power coupling relationship are adjusted, the upward and downward adjustment differences change accordingly, thereby driving the comprehensive flexibility capacity value to adaptively update in a monotonic direction, ensuring that the entire simulation quantization process converges rapidly to the desired accuracy within a finite number of iterations.

[0059] The feasibility of this model strictly depends on the basic operational constraints of power grid dispatch, namely, the actual power value of all adjustable equipment at each time section must always be between the rated upper limit and the rated lower limit, ensuring that the upward and downward adjustment differences are always non-negative values, which can effectively support the goal of high accuracy and high efficiency in the quantitative assessment of the flexibility of power generation, grid, load and storage.

[0060] The beneficial effects are as follows: the bidirectional adjustment potential of equipment is characterized by multiplying the upward and downward power difference, and the comprehensive flexibility capacity value is obtained by superimposing a regularization constant for normalization calculation. This allows for the simultaneous quantification of the bidirectional power adjustment capability of adjustable equipment at the same time. The index solution can be completed solely based on joint simulation data and the inherent rated power parameters of the equipment, without the need for additional external detection and acquisition equipment. The regularization constant is set based on the historical simulation conditions of the system, which can prevent the problem of calculation interruption caused by the simultaneous zeroing of two types of power differences, ensuring the complete generation of adjustment indicators for all time sections throughout the time series. The comprehensive flexibility capacity sequence obtained by arranging the comprehensive flexibility capacity values ​​in time series is directly used as the adjustment index of adjustable resources. The unified quantification standard is adapted to all adjustable equipment on the source side, load side, and storage side, enabling horizontal comparison of the adjustment potential of different types of adjustable resources. The quantification results closely match the joint simulation data produced by the previous scenario sampling and power coupling modeling. The entire calculation process has a continuous data link without additional data conversion loss, improving the calculation stability and result consistency of the source-grid-load-storage adjustable resource flexibility assessment.

[0061] III. Determine the adaptive weight of the corresponding index by the information entropy of the adjustment index value, and sum the adjustment index values ​​according to the adaptive weight to obtain the comprehensive quantization result of the source-grid-load-storage combination, thus completing one simulation quantization process; In this embodiment of the invention, the step of determining the adaptive weight of the corresponding index through the information entropy of the adjustment index value, and then weighting and summing the adjustment index values ​​according to the adaptive weight to obtain the comprehensive quantization result of the source-grid-load-storage combination, thus completing one simulation quantization process, includes: The numerical sequence is divided into several equally wide numerical intervals according to the numerical distribution range in the numerical sequence corresponding to the adjustment index, and the number of values ​​falling into each numerical interval is counted to obtain the frequency distribution vector corresponding to each adjustment index. Each frequency in the frequency distribution vector is compared with the total frequency of the corresponding adjustment index to obtain the distribution probability vector of the adjustment index, and the ratio of the number of non-zero components in the distribution probability vector to the total number of components is used as the discreteness characterization value of the adjustment index. The dispersion characterization value is used as the weight base value of the adjustment index, and the ratio of the weight base value of the adjustment index to the sum of the weight base values ​​of all adjustment indices is used as the adaptive weight of the adjustment index. The product of the adjustment index and the corresponding adaptive weight is used as the weighted component of each adjustment index at each time section, and the sum of the weighted components is used as the comprehensive quantitative value of each time section. The comprehensive quantization values ​​of the cross-sections under all time sections are arranged in chronological order to obtain the comprehensive quantization result of the source-grid-load-storage combination, thus completing one simulation quantization process.

[0062] Extract the complete numerical sequence specific to each adjustment indicator. Define the complete numerical distribution range by the minimum and maximum values ​​of all values ​​within the numerical sequence. Divide the complete numerical distribution range into equal segments using a rule of equal interval spans. The number of segments is a fixed value preset by the system. Determine the segment interval to which each independent value in the numerical sequence belongs. Count the total number of values ​​contained in each segment interval. Organize and collect the statistical counts of values ​​corresponding to all intervals according to the interval arrangement order to form the frequency distribution vector of the corresponding adjustment indicator.

[0063] Retrieve the frequency distribution vector generated for a single regulation indicator, sum the number of values ​​corresponding to all intervals within the vector to obtain the total frequency of this regulation indicator, extract the number of independent values ​​corresponding to each interval within the frequency distribution vector, divide the number of values ​​in a single interval by the total frequency to obtain the probability percentage value corresponding to the single interval, integrate and collect the probability percentage values ​​corresponding to all intervals according to the original arrangement of the intervals to form the distribution probability vector corresponding to this regulation indicator, count the total number of non-zero components within the distribution probability vector, count the total number of all components contained in the distribution probability vector, divide the total number of non-zero components by the total number of all components to obtain the dispersion characterization value of this regulation indicator.

[0064] The dispersion value calculated from a single adjustment indicator is directly set as the weight base value of this adjustment indicator. The weight base values ​​corresponding to all adjustment indicators are summed and the values ​​are added together to obtain the total weight base values. The weight base value of a single adjustment indicator is divided by the total weight base values ​​of all adjustment indicators to obtain the adaptive weight of the corresponding adjustment indicator after the ratio calculation.

[0065] Using a single fixed time segment as an independent accounting unit, the indicator value corresponding to each adjustment indicator under that time segment is extracted and matched with the adaptive weight value generated by matching the adjustment indicator. The two sets of values ​​are multiplied accordingly to obtain the weighted component of the adjustment indicator under that time segment. All weighted components corresponding to all adjustment indicators under the same time segment are collected and the values ​​are added together. After the addition is completed, the comprehensive quantitative value of the segment corresponding to that time segment is obtained.

[0066] The entire simulation data covers all fixed time segments. The comprehensive quantitative values ​​of each time segment are calculated separately and then arranged and integrated in an orderly manner according to the sequential progression of the time segments. After the time sequence integration is completed, a comprehensive quantitative result of the source-grid-load-storage combination that fully covers the entire time sequence data is formed. Once this set of comprehensive quantitative results is generated, the single complete simulation quantitative process is completed.

[0067] The beneficial effects are as follows: by dividing the frequency distribution vector of statistical adjustment index values ​​into equal-width intervals, and relying on the distribution probability vector to calculate the dispersion characterization value, adaptive weights are generated entirely based on the data distribution characteristics of the indicators themselves, eliminating the need for manually preset fixed weight values ​​and the subjective bias caused by manual assignment. Adaptive weights are calculated by converting the dispersion characterization value; adjustment indicators with more discrete data intervals are assigned higher weights, accurately matching the information carrying capacity of the indicator data and improving the distinguishability of quantitative evaluation. The weighted components are obtained by multiplying each adjustment index by its corresponding adaptive weight, and then summed to obtain the cross-sectional comprehensive quantitative value. Completing the weighted calculation in time series can completely preserve the details of the equipment's adjustment capacity at each moment, avoiding the problem of losing time series characteristics in globally uniform weighting. Integrating all cross-sectional comprehensive quantitative values ​​in time series yields the comprehensive quantitative result. The entire process reuses the previously generated adjustment index sequence, the data link is continuous with no additional acquisition cost, and it has strong reproducibility. The final quantitative result can objectively and evenly reflect the comprehensive operational performance of the source-grid-load-storage combination of adjustable resources throughout the entire time period.

[0068] IV. Based on the deviation between the comprehensive quantization result and the preset expected value, the statistical characteristic parameters in the power coupling relationship are reverse-corrected to obtain the corrected statistical characteristic parameters; In this embodiment of the invention, the step of determining the deviation between the comprehensive quantification result and the preset expected value includes: The initial deviation value of the source-grid-load-storage combination under each time section is obtained by subtracting the comprehensive quantitative value of each time section from the expected threshold of the corresponding time section in the preset expected value. The initial deviation values ​​are classified and labeled according to their positive or negative signs. The initial deviation values ​​with positive values ​​are labeled as positive deviation values, and the initial deviation values ​​with negative values ​​are labeled as negative deviation values. The positive and negative deviation values ​​at each time section are arranged in chronological order to obtain a deviation value sequence.

[0069] The reverse correction of the statistical characteristic parameters in the power coupling relationship to obtain the corrected statistical characteristic parameters includes: A positive deviation window is determined based on the continuous distribution interval of the positive deviation value over time, and a negative deviation window is determined based on the continuous distribution interval of the negative deviation value over time. The statistical characteristic parameters corresponding to the positive deviation window are read from the power coupling relationship as the set of parameters to be reduced, and the statistical characteristic parameters corresponding to each time section in the negative deviation window are read from the power coupling relationship as the set of parameters to be increased. The statistical characteristic parameters in the set of parameters to be reduced are compressed along the direction of the preset lower threshold to obtain the set of down-adjustment correction parameters for the positive deviation window; The statistical characteristic parameters in the set of parameters to be increased are stretched along the preset upper threshold direction to obtain the set of adjustment correction parameters for the negative deviation window; The down-modulation correction parameter set and the up-modulation correction parameter set are re-filled into the power coupling relationship according to their respective time sections to obtain the corrected statistical characteristic parameters in the power coupling relationship.

[0070] Retrieve the comprehensive quantitative values ​​of all time segments, retrieve the preset expected values ​​that the system has pre-entered and stored according to the source-grid-load-storage operation standards, extract the expected thresholds that completely match the current segment time identifier within the preset expected values, use a single fixed time segment as an independent accounting unit, subtract the expected threshold matching the same time segment from the comprehensive quantitative value of the segment using the time segment, and obtain the initial deviation value of the source-grid-load-storage combination under the corresponding time segment after the difference accounting operation is completed.

[0071] For the initial deviation values ​​calculated across all time sections, a sign determination and classification labeling operation is performed. The comparison between the numerical value and the number zero is used as the sole classification criterion. Initial deviation values ​​with a value greater than zero are uniformly assigned a fixed label and recorded as positive deviation values. Initial deviation values ​​with a value less than zero are uniformly assigned another fixed label and recorded as negative deviation values. Initial deviation values ​​with a value equal to zero do not belong to either labeling category and are recorded separately.

[0072] After each time segment is classified and labeled, the positive or negative deviation value is extracted. Strictly following the fixed order of advancing the time segments, the deviation values ​​corresponding to the labels of each time segment are arranged and integrated in an orderly manner. After all the time series deviation values ​​are arranged in a regular manner, a deviation value sequence that completely covers all time series nodes is formed.

[0073] Traverse the complete sequence of deviation values, and define an independent time series coverage range by a set of adjacent time segments that are continuously arranged and uniformly marked as positive deviation values. This continuous time series range is uniformly defined as a positive deviation window. Define an independent time series coverage range by a set of adjacent time segments that are continuously arranged and uniformly marked as negative deviation values. This continuous time series range is uniformly defined as a negative deviation window. The window division is determined solely by the sign of the deviation value and the continuity of the time series.

[0074] Retrieve the pre-built power coupling relationship dataset, locate all statistical feature parameters that completely match the time series identifiers within the dataset and all time segments contained in the positive deviation window, and collect and integrate all the matched and filtered statistical feature parameters to form a complete set of parameters to be reduced. Retrieve the power coupling relationship dataset again, locate all statistical feature parameters that completely match the time series identifiers within the dataset and all time segments contained in the negative deviation window, and collect and integrate all the matched and filtered statistical feature parameters to form a complete set of parameters to be increased.

[0075] The retrieval system pre-sets fixed lower threshold values ​​for statistical characteristic parameters. These lower threshold values ​​are defined by fixed boundary values ​​based on the long-term operation statistical intervals of various equipment in the source, grid, load, and storage systems. The system performs numerical compression and adjustment operations on each independent statistical characteristic parameter within the parameter set to be reduced. The adjustment direction is fixed towards the pre-set lower threshold value. After all parameters have been compressed and adjusted, all adjusted parameters are integrated to form a set of reduction correction parameters for a positive deviation window.

[0076] The retrieval system pre-sets fixed upper limit thresholds for statistical characteristic parameters. These upper limit thresholds are defined by fixed boundary values ​​based on the long-term operation statistical intervals of various equipment in the source, grid, load, and storage systems. The system performs numerical stretching adjustment operations on each independent statistical characteristic parameter within the parameter set to be adjusted. The direction of adjustment is fixed towards the pre-set upper limit threshold. After all parameters have been stretched and adjusted, the adjusted parameters are integrated to form a set of adjustment correction parameters for the negative deviation window.

[0077] According to the positive deviation window time section identifiers corresponding to each parameter in the down-adjustment correction parameter set, all down-adjustment correction parameters are backfilled to the same time position within the power coupling relationship, one by one. According to the negative deviation window time section identifiers corresponding to each parameter in the up-adjustment correction parameter set, all up-adjustment correction parameters are backfilled to the same time position within the power coupling relationship, one by one. After completing the replacement and backfilling operation of all time position parameters, the updated parameter content within the power coupling relationship is the corrected statistical characteristic parameter in the power coupling relationship.

[0078] The beneficial effects are as follows: By dividing the corresponding deviation windows based on the continuously distributed positive and negative deviation values ​​in the time series, the complete time series intervals where quantification deviations occur are accurately located, avoiding fragmented parameter adjustments caused by scattered single-point corrections. Based on the window time series, sets of parameters to be reduced and sets of parameters to be increased are extracted from the power coupling relationship, enabling targeted screening of parameters corresponding to the deviation interval. Adjustments are made only to abnormal time series parameters, without modifying the entire time series data, thus reducing the overall workload of data correction. Parameter compression and stretching adjustments are performed along preset fixed upper and lower thresholds. These thresholds are pre-determined based on the historical operating conditions of the equipment, uniformly constraining the parameter adjustment boundaries and preventing parameters from exceeding the feasible operating range of the equipment. The corrected parameters are then backfilled into the original power coupling relationship according to the time series, fully preserving the original data storage architecture and associated indexes. There is no need to reconstruct the dataset, and the corrected statistical characteristic parameters can be directly used in the next round of scenario sampling simulation, forming a closed-loop iterative optimization logic. This continuously reduces the deviation between the comprehensive quantification results and the preset expected values, improving the degree to which the source-grid-load-storage simulation model matches the actual power grid operating conditions.

[0079] V. Based on the corrected statistical characteristic parameters, iterate the simulation quantization process until the comprehensive quantization result converges, and output the comprehensive quantization result at the convergence point.

[0080] In this embodiment of the invention, the step of iterating the simulation quantization process based on the corrected statistical characteristic parameters until the comprehensive quantization result converges, and then outputting the comprehensive quantization result at the point of convergence, includes: The corrected statistical characteristic parameters are substituted into the simulation quantization process for iteration to obtain the updated comprehensive quantization result. The updated comprehensive quantization result is subtracted from the comprehensive quantization result obtained in the previous iteration at the same time section, and the absolute value is taken to obtain the fluctuation amplitude of the source-grid-load-storage combination at each time section. When the fluctuation amplitude at any time segment exceeds the preset convergence tolerance, the updated comprehensive quantization result is used as the comparison benchmark for the next iteration, and the reverse correction operation is re-executed until the fluctuation amplitude at all time segments meets the condition of being less than or equal to the convergence tolerance. When the fluctuation amplitude under all time sections is less than or equal to the convergence tolerance, it is determined that the updated comprehensive quantization result has reached the convergence state, and the updated comprehensive quantization result is marked as the comprehensive quantization result at the time of convergence. The convergence results are arranged according to time sections to obtain the final convergence result, which is then output.

[0081] Retrieve the corrected statistical characteristic parameters after all backfilling and updating operations are completed, and input the set of parameters into the starting data input terminal of the entire simulation quantization process. Fully reproduce the entire simulation quantization operation, including scenario sampling, joint simulation data generation, adjustment index calculation, and adaptive weighted summation. After the entire process is completed, the updated comprehensive quantization result is generated.

[0082] The updated comprehensive quantization results from this iteration are retrieved and compared with the comprehensive quantization results retained from the previous iteration. A single time segment with completely identical time sequence identifiers in the two sets of quantization results is used as an independent comparison unit. The comprehensive quantization values ​​of the two sets of segments corresponding to the same time segment are calculated by difference. The relationship between the value obtained by the difference calculation and the number zero is determined. The absolute value that does not distinguish between positive and negative signs is extracted. After completing the extraction of values ​​from all time segments, the fluctuation amplitude of the source-grid-load-storage combination under each time segment is obtained.

[0083] The system retrieves the convergence tolerance that is fixed in advance according to the simulation accuracy requirements. The fluctuation amplitude of a single time node is compared with the convergence tolerance value as the judgment benchmark. The fluctuation amplitude corresponding to all time segments is reviewed one by one. If the fluctuation amplitude of any time segment is greater than the convergence tolerance value, the updated comprehensive quantization result generated in this iteration is retained as the comparison benchmark data for the next iteration process. The entire set of operations, including deviation calculation, deviation window division, and reverse correction of statistical characteristic parameters, is carried out again. The parameter correction and simulation iteration operations are repeated until the fluctuation amplitude values ​​corresponding to all time segments meet the judgment condition of being less than or equal to the convergence tolerance value.

[0084] The only convergence criterion is that the fluctuation amplitude values ​​of all time segments are less than or equal to the convergence tolerance value. When all time-series nodes meet this criterion, the updated comprehensive quantization results generated in this iteration are determined to have entered the convergence state. A unique convergence identifier is added to the updated comprehensive quantization results. After the marking operation is completed, the comprehensive quantization results at the time of convergence are formed.

[0085] Extract all cross-sectional comprehensive quantization values ​​from the convergence result that has been marked with convergence, and arrange and integrate them in an orderly manner according to the fixed sequence of time cross-sections. After the time sequence is regularized, the final complete comprehensive quantization result is formed. Perform an external output operation on this set of data to end the entire iterative simulation quantization process.

[0086] The beneficial effects include fully incorporating the corrected statistical characteristic parameters into the simulation quantization process for iterative calculations, reusing the existing complete simulation data link without needing to rebuild the data association structure, thus reducing repetitive modeling work. By calculating the fluctuation amplitude of the two iterations of quantization results on a cross-section basis, and using a preset convergence tolerance as a unified judgment standard, the system accurately identifies numerical fluctuations in the time-series dimension, avoiding the defect of global overall comparison masking single-time-period deviations. When the cross-section fluctuation amplitude exceeds the convergence tolerance, the system automatically performs reverse correction and simulation iteration in a loop, forming a closed-loop self-optimization logic that continuously narrows the gap between the quantization results and the expected operating conditions. Convergence is only determined after all time-series cross-section fluctuation amplitudes meet the threshold constraints, ensuring the stability of the simulation values ​​throughout the entire time period. The converged quantization results are directly output after being normalized according to the time series, and the unified data format facilitates subsequent evaluation of the source-grid-load-storage system operation. The entire iterative mechanism relies on the previous deviation correction module for coordinated operation, improving the matching accuracy between the simulation results and the actual operating state of the power grid, and the scheme has strong reproducibility.

[0087] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0088] Figure 2 The results of the time-series simulation during the system characteristic evaluation phase are shown. The horizontal axis represents the simulation time-series cross-section, and the vertical axis represents the quantitative indicators of the flexibility characteristics of the four types of source-load-storage equipment. The curves for source-side adjustable equipment show a continuous upward trend in the early stages, reaching a peak between cross-sections 4 and 6 before gradually declining. The curves for load-side adjustable equipment show a gentle downward trend throughout, with smaller overall fluctuations. Energy storage equipment A and B exhibit an inverse coupling pattern: energy storage A first rises and then falls, entering a negative range in the middle, while energy storage B first falls to the bottom of the negative control zone and then rises back to the positive range in the latter half, directly reflecting the complementary adjustment characteristics of the three types of resources: source, load, and storage. All four curves deviate significantly from the adaptive convergence tolerance line, and the flexibility indicators of each equipment show obvious alternating positive and negative fluctuations. This indicates that before deviation calibration, the flexibility characteristics of each unit in the system have significant deviations, and the original simulation data cannot meet the quantitative accuracy requirements, thus justifying the necessity of subsequent multiple rounds of deviation calibration.

[0089] Figure 3The time-series curves for stabilizing the source-load-storage flexibility index during the feature extraction stage correspond to the time-series simulation results of the algorithm's feature extraction and initial evaluation phases. The vertical dashed line represents the stage transition boundary, with the left section representing the pre-feature extraction period and the right section representing the formal evaluation interval. To the left of the boundary line, the indices for the four types of equipment (source, load, and storage) fluctuate dramatically. Energy storage device B drops to the negative control range at its lowest point, while energy storage device A, source-side adjustable devices, and load-side devices all exhibit significant peak abrupt changes. After crossing the boundary line and entering the evaluation phase, the flexibility indices of the four types of equipment quickly converge to a narrow fluctuation range. The load-side adjustable devices enter a long-term stable operating state after section -35, with a significantly narrowed curve fluctuation amplitude. The source-side and two energy storage device curves exhibit small, gentle fluctuations synchronously, without significantly deviating from the positive control region throughout the process. This clearly demonstrates that after the feature extraction stage completes the solidification of the system's basic parameters, the flexibility quantification indices of various adjustable devices can quickly escape initial disturbances and enter a stable observation range, providing stable benchmark data for subsequent deviation tracking.

[0090] Figure 4 The curve shows the tracking time series curve for the load-side negative deviation triggering compressed calibration. After evaluation, the curve represents the time series simulation curve for the deviation tracking phase. The vertical dashed line divides the initial stable evaluation period and the deviation tracking correction period. To the left of the dashed line, the indicators of the four types of equipment generally maintain a small positive fluctuation, indicating stable operation. After crossing the boundary line, the load-side adjustable equipment indicators rapidly drop into the pink negative deviation control zone, significantly deviating from the system's expected quantification value. This triggers the compressed calibration mechanism, with the two energy storage devices and the source-side equipment synchronously following the downward trend. Multiple curves simultaneously show negative offset fluctuations, reaching an extreme value at section -53. Subsequently, the indicators of each device slowly recover, returning to the positive control range for some periods. This curve fully reproduces the deviation identification triggering logic, intuitively verifying that when the load-side flexibility indicators deviate from the preset expected value, the system can accurately capture the time series deviation and initiate the compressed calibration process, clearly demonstrating the entire process of identifying and quantifying errors and triggering correction actions in the deviation tracking phase.

[0091] Figure 5The time series curves show the entire convergence process of flexibility index under multi-round iterative calibration, illustrating the convergence process of multi-round simulated iterative calibration after deviation identification. The vertical dashed line separates the original deviation tracking section from the multi-round calibration iteration interval. In the original time series to the left of the dashed line, the load-side equipment is in the negative deviation range for a long time, and the overall index is low. After entering the first round of the calibration cycle, the flexibility index of the four types of equipment simultaneously rises to the positive stretching control zone, with the energy storage equipment B having the highest peak. Subsequently, with the increase of calibration rounds, the four curves synchronously and uniformly decay, gradually approaching the purple adaptive convergence tolerance line (ε≤0.02). In the later calibration section, the fluctuation amplitude of all equipment indexes is close to the convergence threshold, and the curve fluctuations are slight. This time series change fully demonstrates that the proposed multi-round calibration iterative algorithm has efficient convergence characteristics, which can gradually offset the quantization negative deviation brought by the load side in the early stage through multi-round stretching correction. After a limited number of iterations, the flexibility index of all types of adjustable equipment can meet the convergence accuracy requirements, verifying the quantization accuracy optimization effect of the closed-loop correction mechanism.

[0092] Figure 6 The simulation curves for the multi-round deviation calibration iteration process are shown. The dashed line on the left is the boundary between the evaluation and calibration stages, the horizontal axis is the calibration iteration section, and the purple dashed line represents the 0.02 adaptive convergence tolerance. In the early stage of the iteration, the flexibility indices of the four types of equipment all showed a momentary spike, followed by synchronous monotonic decay, and the curves as a whole fell within the positive deviation stretching control zone. As the calibration rounds progressed, the four curves gradually approached the convergence tolerance line. By the 10th iteration, all indices were stable within the threshold range, and the curves in subsequent calibration rounds almost coincided with the tolerance line with no significant fluctuations. This curve intuitively verifies that the proposed adaptive deviation calibration mechanism has a rapid convergence capability, requiring only 10 iterations to eliminate the quantification deviation of the source-load-storage equipment flexibility, proving that this method can efficiently correct the errors in the original simulation data and achieve high-precision unified quantification of the system flexibility indices.

[0093] This application embodiment can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0094] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for quantifying source-grid-load-storage flexibility based on simulation modeling and analysis, characterized in that, include: Ⅰ. Scenario sampling is performed on the statistical characteristic parameters of the standardized sequence in the source-grid-load-storage combination to generate the initial scenario parameters of the standardized sequence, and the initial scenario parameters are mapped to a preset power coupling relationship to obtain the joint simulation data corresponding to the initial scenario parameters; II. Calculate the adjustment index of the adjustable resources corresponding to the joint simulation data; III. Determine the adaptive weight of the corresponding index by the information entropy of the adjustment index value, and sum the adjustment index values ​​according to the adaptive weight to obtain the comprehensive quantization result of the source-grid-load-storage combination, thus completing one simulation quantization process; IV. Based on the deviation between the comprehensive quantization result and the preset expected value, the statistical characteristic parameters in the power coupling relationship are reverse-corrected to obtain the corrected statistical characteristic parameters; V. Based on the corrected statistical characteristic parameters, iterate the simulation quantization process until the comprehensive quantization result converges, and output the comprehensive quantization result at the convergence point.

2. The source-grid-load-storage flexibility quantification method based on simulation modeling analysis as described in claim 1, characterized in that, Before performing scene sampling, the following are included: The ratio of the actual operating power value to the corresponding rated power value of the adjustable equipment in the source-grid-load-storage combination is used as the standardized power value of the adjustable equipment at each time section, and the standardized power values ​​are arranged in time sequence to obtain the standardized power sequence corresponding to the adjustable equipment. The standardized power sequences are grouped and merged according to equipment type to obtain the standardized equipment sequence of the source-grid-load-storage combination. Sliding window truncation is performed on the standardized sequence of the device, and the arithmetic mean, variance and kurtosis of the standardized sequence in each sliding window are used as the statistical feature sub-parameters of the sliding window; The statistical feature sub-parameters corresponding to the same equipment type are concatenated according to the window time sequence to obtain the statistical feature parameters corresponding to each type of equipment.

3. The source-grid-load-storage flexibility quantification method based on simulation modeling analysis as described in claim 2, characterized in that, The step of sampling statistical characteristic parameters of the standardized sequence in the source-grid-load-storage combination for scenarios to generate initial scenario parameters of the standardized sequence, and mapping the initial scenario parameters to a preset power coupling relationship to obtain joint simulation data corresponding to the initial scenario parameters, includes: The mean sequence, variance sequence, and kurtosis sequence of the equipment type are read from the statistical feature parameters. The values ​​at each time point in the mean sequence are used as the midline of the interval, the values ​​at each time point in the variance sequence are used as the basic radius length, and the values ​​at each time point in the kurtosis sequence are used as the boundary correction coefficients to construct the sampling interval of the equipment type. Within the sampling interval, stratified equidistant sampling is performed on the device type to obtain the initial scene sub-parameters of the device type at each time point; The initial scene sub-parameters are merged according to device type identifier and time identifier to obtain the initial scene parameters of the standardized sequence; Based on the initial scenario parameters, the reference power mapping value of each device type at each time is found from the preset power coupling relationship, and the power offset corresponding to the sampled value in the initial scenario parameters is superimposed on the reference power mapping value to obtain the power mapping value of the device type. By merging the power mapping values ​​of all device types at each time point, the joint simulation data corresponding to the initial scenario parameters is obtained.

4. The source-grid-load-storage flexibility quantification method based on simulation modeling analysis as described in claim 3, characterized in that, The step of finding the reference power mapping value for each device type at each time from the preset power coupling relationship includes: Adjustable devices with values ​​of source-side, load-side, and storage-side respectively are selected from the device type field of the adjustable devices and designated as source-side devices, load-side devices, and storage devices, respectively. The access nodes of the source-side device, load-side device, and storage-side device are respectively marked as source node, load node, and storage node, and the node number of each associated storage node is bound with the corresponding device type identifier to generate the node device association index of the adjustable device; Using the source node, the load node, and the storage node as vertices of the graph, and the transmission lines and transformers between the nodes as edges of the graph, an undirected topology graph reflecting the node connection relationships is constructed. First, using the source node as the starting vertex and the load node or the storage node as the target vertex, then using the storage node as the starting vertex and the load node as the target vertex, perform a breadth-first traversal on the undirected topology graph to search for all connected paths between the starting vertex and the target vertex.

5. The source-grid-load-storage flexibility quantification method based on simulation modeling analysis as described in claim 4, characterized in that, The step of finding the reference power mapping value for each device type at each time from the preset power coupling relationship includes: The connected path with the source node as the starting vertex and the load node as the target vertex is recorded as the power transmission path; the connected path with the storage node as the target vertex is recorded as the power charging and discharging path; and the connected path with the storage node as the starting vertex is recorded as the power feedback path. Assign a path identifier to each connected path found, and bind the path identifier to the nodes at both ends of the power transmission path, the power charging and discharging path, and the power feedback path to generate a path node association index for each path; Based on the rated voltage level of each node, the maximum transmission capacity and transmission loss rate are configured for each path, and the path parameter record set is written according to the path identifier of each path to generate the path parameter index of each path. The statistical feature parameters are associated with the node device association index, the path node association index, and the path parameter index, and compiled to generate a power reference mapping value for each device type identifier. A power coupling relationship is constructed using the device type identifier, node number, path identifier, and time segment as the joint primary key, and the power reference mapping value as the attribute value.

6. The source-grid-load-storage flexibility quantification method based on simulation modeling analysis as described in claim 1, characterized in that, The calculation of the adjustment index for the adjustable resources corresponding to the co-simulation data includes: Read the actual power values ​​of the adjustable device at each time section from the joint simulation data, and read the upper limit and lower limit of the rated power of the adjustable device. The difference between the actual power value and the corresponding upper limit of rated power and lower limit of rated power is obtained by subtracting them from the actual power value, respectively, to obtain the power adjustment difference and the power adjustment difference at each time section. The comprehensive flexibility capacity value of the adjustable equipment at each time segment is calculated based on the upward power adjustment difference and the downward power adjustment difference. The formula for calculating the comprehensive flexibility capacity value is as follows: ; in, Indicates the first The adjustable device in the first The comprehensive flexibility capacity value under a specific time segment. Indicates the first The adjustable device in the first The difference in upward power across different time segments Indicates the first The adjustable device in the first The difference in downward power across different time segments This represents the preset regularization constant; The comprehensive flexibility capacity values ​​are arranged in chronological order to generate a comprehensive flexibility capacity sequence for each adjustable device, and the comprehensive flexibility capacity sequence is used as an adjustment index for the adjustable resources corresponding to the joint simulation data.

7. The source-grid-load-storage flexibility quantification method based on simulation modeling analysis as described in claim 1, characterized in that, The process of determining the adaptive weight of the corresponding index through the information entropy of the adjustment index value, and then weighting and summing the adjustment index values ​​according to the adaptive weight to obtain the comprehensive quantization result of the source-grid-load-storage combination, completes one simulation quantization process, including: The numerical sequence is divided into several equally wide numerical intervals according to the numerical distribution range in the numerical sequence corresponding to the adjustment index, and the number of values ​​falling into each numerical interval is counted to obtain the frequency distribution vector corresponding to each adjustment index. Each frequency in the frequency distribution vector is compared with the total frequency of the corresponding adjustment index to obtain the distribution probability vector of the adjustment index, and the ratio of the number of non-zero components in the distribution probability vector to the total number of components is used as the discreteness characterization value of the adjustment index. The dispersion characterization value is used as the weight base value of the adjustment index, and the ratio of the weight base value of the adjustment index to the sum of the weight base values ​​of all adjustment indices is used as the adaptive weight of the adjustment index. The product of the adjustment index and the corresponding adaptive weight is used as the weighted component of each adjustment index at each time section, and the sum of the weighted components is used as the comprehensive quantitative value of each time section. The comprehensive quantization values ​​of the cross-sections under all time sections are arranged in chronological order to obtain the comprehensive quantization result of the source-grid-load-storage combination, thus completing one simulation quantization process.

8. The source-grid-load-storage flexibility quantification method based on simulation modeling analysis as described in claim 1, characterized in that, The deviation between the comprehensive quantification result and the preset expected value includes: The initial deviation value of the source-grid-load-storage combination under each time section is obtained by subtracting the comprehensive quantitative value of each time section from the expected threshold of the corresponding time section in the preset expected value. The initial deviation values ​​are classified and labeled according to their positive or negative signs. The initial deviation values ​​with positive values ​​are labeled as positive deviation values, and the initial deviation values ​​with negative values ​​are labeled as negative deviation values. The positive and negative deviation values ​​at each time section are arranged in chronological order to obtain a deviation value sequence.

9. The source-grid-load-storage flexibility quantification method based on simulation modeling analysis as described in claim 8, characterized in that, The reverse correction of the statistical characteristic parameters in the power coupling relationship to obtain the corrected statistical characteristic parameters includes: A positive deviation window is determined based on the continuous distribution interval of the positive deviation value over time, and a negative deviation window is determined based on the continuous distribution interval of the negative deviation value over time. The statistical characteristic parameters corresponding to the positive deviation window are read from the power coupling relationship as the set of parameters to be reduced, and the statistical characteristic parameters corresponding to each time section in the negative deviation window are read from the power coupling relationship as the set of parameters to be increased. The statistical characteristic parameters in the set of parameters to be reduced are compressed along the direction of the preset lower threshold to obtain the set of down-adjustment correction parameters for the positive deviation window; The statistical characteristic parameters in the set of parameters to be increased are stretched along the preset upper threshold direction to obtain the set of adjustment correction parameters for the negative deviation window; The down-modulation correction parameter set and the up-modulation correction parameter set are re-filled into the power coupling relationship according to their respective time sections to obtain the corrected statistical characteristic parameters in the power coupling relationship.

10. The source-grid-load-storage flexibility quantification method based on simulation modeling analysis as described in claim 1, characterized in that, The simulation quantization process is iterated based on the corrected statistical characteristic parameters until the comprehensive quantization result converges. The comprehensive quantization result at the point of convergence is then output, including: The corrected statistical characteristic parameters are substituted into the simulation quantization process for iteration to obtain the updated comprehensive quantization result. The updated comprehensive quantization result is subtracted from the comprehensive quantization result obtained in the previous iteration at the same time section, and the absolute value is taken to obtain the fluctuation amplitude of the source-grid-load-storage combination at each time section. When the fluctuation amplitude at any time segment exceeds the preset convergence tolerance, the updated comprehensive quantization result is used as the comparison benchmark for the next iteration, and the reverse correction operation is re-executed until the fluctuation amplitude at all time segments meets the condition of being less than or equal to the convergence tolerance. When the fluctuation amplitude under all time sections is less than or equal to the convergence tolerance, it is determined that the updated comprehensive quantization result has reached the convergence state, and the updated comprehensive quantization result is marked as the comprehensive quantization result at the time of convergence. The convergence results are arranged according to time sections to obtain the final convergence result, which is then output.