A method and system for adjustable resource flexibility evaluation based on adjusted capacity

Abstract

The present application relates to a kind of adjustable resource flexibility evaluation method and system based on regulating capacity, belong to the field of adjustable resource evaluation and energy management information processing of power system.It includes obtaining the hourly output sequence of adjustable resource and the hourly supply-demand gap sequence of target power grid;Correlation coefficient of hourly output sequence is calculated;Based on resource complementary directed graph, the betweenness centrality of each node is calculated;Extract maximum regulating capacity, maximum power ramp rate and longest duration of regulation;Calculate the time average of first-order difference sequence, generate option pricing eigenvalue;Based on option pricing eigenvalue, filter nodes and summarize maximum regulating capacity.The present application adopts the mode of supply-demand gap period identification, resource complementary graph analysis and node eigenvalue calculation, can generate the effective regulating capacity evaluation result of adjustable resource, improve the pertinence and distinction of adjustable resource flexibility evaluation.

Description

Technical Field

[0001] This invention relates to the field of power system adjustable resource assessment and energy management information processing, and particularly to a method and system for assessing the flexibility of adjustable resources based on adjustable capacity. Background Technology

[0002] With the rapid development of new power systems and virtual power plants, distributed energy storage, photovoltaics, and various adjustable loads are being connected to the grid in large quantities. In order to achieve optimal resource allocation, operation management, and macro-dispatch decisions in the power market environment, a systematic flexibility assessment of massive distributed adjustable resources is required.

[0003] In related technologies, Chinese invention patent CN115018362A discloses a method and apparatus for assessing regional resource regulation capacity under source-load interaction scenarios. This method acquires adjustable resource data within a region, generates a source-load interaction scenario set for power system resource regulation based on the adjustable resource data, obtains a predicted load curve based on the source-load interaction scenario set, corrects the predicted load curve using a transferable load model to obtain a second load curve, calculates the ratio of load power to adjustable power based on the predicted load curve and the second load curve to obtain the first power system's adjustable capacity, and finally quantitatively evaluates the first power system's adjustable capacity based on regulation performance indicators such as the number of historical load aggregator adjustments and the historical total adjusted power.

[0004] The aforementioned technologies primarily rely on static comparisons of electricity ratios for simple capacity calculations, treating each resource as an isolated entity and evaluating it based on historical conventional indicators. This approach fails to accurately measure the true regulatory value of multi-source resources under extreme supply and demand conditions, nor does it reflect the complementary attributes of resources under complex operating conditions. Consequently, the resulting flexibility assessment is dimensionally singular and has low accuracy, making it difficult to meet the in-depth decision-making needs of modern electricity market transactions and macro-level system management. Summary of the Invention

[0005] To address the aforementioned technical problems, or at least partially address them, and to overcome the aforementioned shortcomings, this invention provides a method and system for assessing the flexibility of adjustable resources based on adjustable capacity.

[0006] In a first aspect, the present invention provides a method for assessing the flexibility of adjustable resources based on adjustable capacity, comprising: Obtain the hourly output sequence of adjustable resources and the hourly supply-demand gap sequence of the target power grid; Based on the first-order difference sequence and standard deviation of the hourly supply and demand gap sequence, a continuous time segment is extracted and aggregated to generate a time window set; Within the time window set, the correlation coefficient of the hourly output sequence is calculated. If the correlation coefficient is negative, the absolute value is taken as the edge weight, and a resource complementary directed graph is constructed with the adjustable resources as nodes. Based on the resource complementarity directed graph, the betweenness centrality of each node is calculated, and the hub weight coefficient is generated according to the betweenness centrality; within the time window set, the maximum regulation capacity, the maximum power ramp rate, and the longest continuous regulation duration are extracted, and the comprehensive regulation capacity coefficient is calculated; within the time window set, the time mean of the first-order difference sequence is calculated. Option pricing feature values ​​are generated based on the maximum adjustment capacity, the time average, the hub weight coefficient, and the comprehensive adjustment capacity coefficient. The effective regulation capacity assessment value of the target power grid is calculated by summarizing the maximum regulation capacity by selecting effective nodes based on option pricing characteristic values ​​and screening them.

[0007] Furthermore, the aggregated generation time window set includes: Raw metering data is collected for adjustable resources, and load deviation data is collected for the target power grid. Interpolation and alignment are performed at a uniform timestamp granularity to generate hourly output sequences and hourly supply-demand gap sequences. The root mean square deviation of the hourly supply and demand gap sequence is calculated as the standard deviation of the hourly supply and demand gap sequence; A first-order difference sequence is generated by subtracting the preceding term from the following term in the hourly supply-demand gap sequence. Based on the first-order difference sequence and standard deviation of the hourly supply and demand gap sequence, a continuous time segment is extracted and aggregated to generate a time window set.

[0008] Furthermore, the aggregated generation time window set includes: Extract target data points from the first-order difference sequence that have positive values ​​and are greater than the standard deviation, and extract the time index of the target data points; The time indices are concatenated to generate continuous time segments. These continuous time segments are then stored sequentially in a list data structure to generate a time window set. All extracted time indices are sorted in ascending order by time. The time span difference between adjacent time indices is calculated. If the span difference between two adjacent time indices equals the base step size of the unified timestamp granularity, these two time indices are considered to be physically continuous. Boundary aggregation is performed on all continuous time indices, and a closed time interval is constructed using the start and end time indices of the continuous sequence as the continuous time segment.

[0009] Furthermore, the construction of the resource-complementary directed graph includes: Within the time window set, calculate the covariance of the hourly output sequence of any two adjustable resources and the standard deviation of the hourly output sequence of any two adjustable resources, and divide the covariance by the product of the two standard deviations to generate the Pearson correlation coefficient. If the Pearson correlation coefficient is less than zero, it is determined that the two corresponding adjustable resources have complementary attributes. An absolute value transformation is performed to generate edge weights. The adjustable resources are instantiated as graph structure nodes, and the graph structure nodes are connected with the edge weights as connection attributes to construct a resource complementary directed graph.

[0010] Furthermore, the generation of hub weight coefficients based on betweenness centrality includes: Based on the resource-complementary directed graph, for any node being evaluated in the network, any two other source nodes and target nodes in the network that do not contain the node being evaluated are addressed. The total number of shortest paths between the source node and the target node is calculated, and the frequency of the shortest path passing through the current node is counted in the total number of shortest paths. The frequency of the shortest path is divided by the total number of shortest paths to generate the betweenness centrality of the node being evaluated. The betweenness centrality is extracted to construct a centrality value sequence. The element with the largest value is extracted from the centrality value sequence to generate the maximum betweenness centrality value. The betweenness centrality is divided by the maximum betweenness centrality value to generate the hub weight coefficient.

[0011] Furthermore, breadth-first search or Dijkstra's algorithm can be used to calculate the total number of shortest paths between the source node and the target node.

[0012] Furthermore, generating the comprehensive coefficient of adjustment capability includes: Within the time window set, the operating data of each node is extracted to parse out the maximum regulation capacity, maximum power ramp rate and longest continuous regulation duration, and the feature value sequence is constructed by category; A maximum value retrieval operation is performed on each of the feature value sequences to generate a global maximum value. Each feature value sequence is then divided by the global maximum value as a dividend to generate a quotient value. Finally, a multiplication operation is performed on the quotient value to generate a comprehensive coefficient for adjustment capability.

[0013] Furthermore, the generated option pricing feature values ​​include: Within the time window set, all numerical items of the first-order difference sequence are extracted, and a summation operation is performed on all the numerical items to generate a total value. The total value is then divided by the time span of the time window set to generate the time mean. Extract the maximum adjustment capacity, the hub weight coefficient, and the comprehensive adjustment capability coefficient, and construct a quaternary product formula with the time mean. Perform scalar multiplication algebra operation to generate option pricing feature values.

[0014] Furthermore, the step of summarizing the maximum regulating capacity to calculate the effective regulating capacity assessment value of the target power grid by screening effective nodes based on option pricing characteristic values ​​includes: Extract the option pricing feature values, perform an accumulation operation to generate a feature sum, and divide the feature sum by the total number of nodes to generate an arithmetic mean; Traverse all nodes and compare the value of the option pricing feature with the arithmetic mean. Mark nodes whose option pricing feature is less than the arithmetic mean as invalid and remove them from the data. Extract the valid nodes that have not been removed, extract the maximum regulating capacity and perform scalar summation operation, and output the effective regulating capacity assessment value of the target power grid.

[0015] Secondly, the present invention provides an adjustable resource flexibility assessment system based on adjustable capacity, comprising: a time window generation module, used to obtain the hourly output sequence of adjustable resources and the hourly supply-demand gap sequence of the target power grid, and to extract continuous time segments based on the first-order difference sequence and standard deviation of the hourly supply-demand gap sequence, and aggregate to generate a time window set; The resource complementarity graph construction module is used to calculate the correlation coefficient of the hourly output sequence within the time window set. If the correlation coefficient is negative, the absolute value is taken as the edge weight, and a resource complementarity directed graph is constructed with the adjustable resources as nodes. The hub weight calculation module is used to calculate the betweenness centrality of each node based on the resource complementarity directed graph, and generate hub weight coefficients based on the betweenness centrality. The regulation capacity calculation module is used to extract the maximum regulation capacity, the maximum power ramp rate, and the longest continuous regulation duration within the time window set, and to calculate the comprehensive regulation capacity coefficient. The option pricing feature value calculation module is used to calculate the time mean of the first difference sequence within the time window set, and generate option pricing feature values ​​based on the maximum adjustment capacity, the time mean, the hub weight coefficient, and the adjustment capacity comprehensive coefficient. The effective regulation capacity assessment module is used to filter effective nodes based on option pricing characteristic values ​​and summarize the maximum regulation capacity to calculate the effective regulation capacity assessment value of the target power grid.

[0016] The technical solutions provided in the embodiments of the present invention have the following advantages compared with the prior art: This invention performs first-order differential processing on the hourly supply and demand gap sequence of the target power grid and extracts continuous time segments with positive values ​​that exceed the corresponding standard deviation to generate a set of time windows. This allows the flexibility assessment to focus on the critical period when the supply and demand gap is rapidly expanding, reduces the dilution of the results by the full-time average processing, and improves the responsiveness of the assessment process to the characteristics of time-period changes.

[0017] Within a time window set, this invention constructs a resource complementarity directed graph based on the hourly output sequence of adjustable resources, and generates hub weight coefficients by combining betweenness centrality. This allows for further characterization of the complementary relationships between resources and differences in node connectivity beyond the adjustment capacity, enabling the evaluation results to simultaneously reflect the individual capabilities of resources and the structural characteristics of resource combinations.

[0018] This invention combines the maximum regulating capacity, time average, hub weight coefficient, and comprehensive regulating capacity coefficient to generate option pricing characteristic values. After screening effective nodes by using the arithmetic mean as the boundary, the maximum regulating capacity is summarized to form the effective regulating capacity assessment value of the target power grid, thus establishing a complete assessment chain from time period identification and structural analysis to capacity summarization. Attached Figure Description

[0019] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 A flowchart illustrating a method for assessing the flexibility of adjustable resources based on adjustable capacity, provided in an embodiment of the present invention; Figure 2 The adjustable resource sequence Pearson correlation coefficient matrix provided in this embodiment of the invention; Figure 3 This is an evaluation graph of node betweenness centrality and hub weight coefficients for a graph structure provided in an embodiment of the present invention. Figure 4 This is a schematic diagram of the structure of an adjustable resource flexibility assessment system based on adjustable capacity, according to an embodiment of the present invention.

[0022] Figure 5 This is a schematic diagram of an adjustable resource flexibility assessment device based on adjustable capacity, provided in an embodiment of the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0025] Example 1 This invention provides a method for assessing the flexibility of adjustable resources based on adjustable capacity. It combines supply-demand gap period identification, resource complementarity graph analysis, and node characteristic value calculation to generate effective adjustable capacity assessment results for adjustable resources, thereby improving the relevance and discriminativeness of adjustable resource flexibility assessment. Figure 1 As shown, the method in this embodiment specifically includes: Obtain the hourly output sequence of adjustable resources and the hourly supply-demand gap sequence of the target power grid.

[0026] Calculate the first-order difference sequence and the standard deviation of the hourly supply and demand gap sequence, extract the continuous time segments in the first-order difference sequence that are positive and greater than the standard deviation of the hourly supply and demand gap sequence, and aggregate them to generate a time window set. Optionally, the aggregation generation time window set includes: Collect raw metering data of adjustable resources, collect target power grid load deviation data, perform interpolation and alignment according to uniform timestamp granularity, and generate hourly output sequence and hourly supply and demand gap sequence; The root mean square deviation of the hourly supply and demand gap sequence is calculated as the standard deviation of the hourly supply and demand gap sequence; During the data preparation and standardization phase, the data acquisition gateway collects raw metering data of various adjustable resources from the underlying communication interface, and simultaneously obtains load deviation data reflecting the real-time imbalance between power generation and load from the dispatch center database. Due to differences in data reporting cycles among different physical devices and business platforms, interpolation and alignment operations are performed on the two types of data according to the set time granularity to eliminate misalignments and gaps in the time dimension, thereby constructing an hourly output sequence and an hourly supply-demand gap sequence with standard time resolution. To objectively quantify the baseline fluctuation level of the supply-demand gap, the root mean square deviation principle from statistics is further introduced to calculate the standard deviation of the hourly supply-demand gap sequence. The calculation formula is as follows: , in, It represents the standard deviation of the hourly supply and demand gap sequence, and its physical meaning is the baseline of the normal load fluctuation of the target power grid within a specific observation period; This represents the total number of data points in the hourly supply and demand gap sequence; This represents the hourly supply-demand gap at time step t. The root mean square information of the sequence fluctuations is extracted using a formula, providing a reliable data-driven threshold for identifying abnormally drastic fluctuations.

[0027] For example, an industrial park includes two types of adjustable resources: energy storage power stations and temperature-controlled air conditioning systems. The bottom-level gateway collects the active power of the energy storage power station and the electricity load of the air conditioning system every 3 minutes as raw metering data; simultaneously, the dispatching-side gateway records the load deviation data for the current area every 5 minutes. Using 5 minutes as a unified timestamp granularity, a linear interpolation algorithm is used to resample and align the raw metering data from the 3-minute intervals, generating hourly output sequences and hourly supply-demand gap sequences with a strict time step of 5 minutes. If 288 data points are extracted in the past 24 hours, these 288 hourly supply-demand gap values ​​are extracted and substituted into the root mean square deviation formula for calculation. Assuming the final calculated value is 15 MW, this 15 MW is identified as the standard deviation of the hourly supply-demand gap sequence, serving as an objective reference line for measuring the degree of supply-demand deterioration.

[0028] A first-order difference sequence is generated by subtracting the preceding term from the following term in the hourly supply-demand gap sequence. In the generation of the time feature set, to capture the dynamic and urgent demand of the power grid for regulation resources, a difference operation is performed on the hourly supply-demand gap sequence along the time axis. The difference operation subtracts the data value of the immediately preceding time node from the data value of the current time node, thus outputting the first-order difference sequence. The physical essence of the difference operation is to calculate the slope of the supply-demand gap change, which directly reflects the rate of gap deterioration. Its calculation formula is as follows: , in, This represents the first-order difference value at time step t; This represents the hourly supply-demand gap at time step t. This represents the hourly supply and demand gap at time step t-1.

[0029] Based on the first-order difference sequence and standard deviation, extract continuous time segments and aggregate them to generate a time window set. Extract target data points in the first-order difference sequence that are positive and greater than the standard deviation, and extract the time index of the target data points; concatenate the time indices to generate continuous time segments, and store the continuous time segments sequentially into a list data structure to generate a time window set. Aggregate the continuous time segments to generate a time window set.

[0030] After obtaining the first-order difference sequence, it is compared with the standard deviation of the hourly supply-demand gap sequence using Boolean logic. When the first-order difference value of a certain time step is greater than zero and also greater than the standard deviation of the hourly supply-demand gap sequence, it indicates that the supply-demand gap at that moment is not only expanding positively, but the rate of expansion significantly exceeds the historical normal fluctuation level. Therefore, this time step is determined to be a high-potential scarcity period. Adjacent time steps on the time axis that all meet the above conditions are spliced ​​together to form a continuous time segment spanning multiple time steps. Finally, multiple independent continuous time segments are stored in a list data structure and encapsulated as a set of time windows, serving as the absolute time boundary for constructing the resource complementarity map.

[0031] For example, in the operational scenario of an industrial park, the standard deviation of the hourly supply-demand gap sequence has been calculated from objective data to be 15 MW. The analysis engine performs a traversal operation of subtracting the preceding term from the subsequent term in the hourly supply-demand gap sequence for the day. At 14:00, the recorded hourly supply-demand gap value is 30 MW, while at 14:05, this value surges to 50 MW. Substituting into the first-order difference formula, it can be seen that the first-order difference value at 14:05 is 50 MW minus 30 MW, resulting in 20 MW. Since 20 MW is greater than zero and greater than the benchmark of 15 MW, it is determined that the extraction condition is met at 14:05. If the calculated first-order difference values ​​at 14:10 and 14:15 also reach 18 MW and 16 MW respectively, these three adjacent 5-minute time steps are concatenated together to define a continuous time segment of 15 minutes in length. All similar abnormal incremental periods that meet the stringent extraction criteria are extracted and stored in a structured array on the same day, thus obtaining the final set of time windows. This ensures that all subsequent calculations of the entire evaluation method focus only on the critical moments when the power grid needs the most urgent assistance.

[0032] Optionally, the aggregation generation time window set includes: Locate the target data point in the first-order difference sequence whose value is greater than zero and greater than the standard deviation of the hourly supply and demand gap sequence, and extract the time index of the target data point; During the data processing for locating periods of high-potential demand adjustment, the entire pre-generated first-order difference sequence is traversed sequentially. To pinpoint critical nodes where the supply and demand situation of the target power grid deteriorates sharply, a dual threshold condition judgment is performed on each data point in the first-order difference sequence. The strict mathematical logic for marking a data point as a target data point is as follows: the first-order difference value of the data point must be positive, and the absolute magnitude of the first-order difference value must be strictly greater than the pre-calculated standard deviation of the hourly supply and demand gap sequence. The physical meaning of this dual judgment is that it not only requires that the supply and demand gap at the current moment is in an expanding state, i.e., the first-order difference is positive, but also requires that the slope of its expansion significantly exceeds the background fluctuation noise of the target power grid during normal operation, i.e., it is greater than the standard deviation. The condition extraction formula is as follows: , in, This represents the set of time indices for the extracted target data points. A time index is the coordinate point of the data on the time axis. This represents the current time index being traversed; Represents the time index The first-order difference value at the point is derived from the result of subtracting the preceding term from the following term for the hourly supply and demand gap sequence; This represents the standard deviation of the hourly supply and demand gap sequence, derived from the static baseline value obtained by calculating the root mean square deviation of the hourly supply and demand gap sequence. All target data points that meet the conditions are extracted using the above logical expression, and their corresponding time indices are recorded.

[0033] For example, suppose the target power grid has 5 consecutive time sampling points within a certain observation period, and the corresponding time indices are denoted as follows: The hourly standard deviation of the supply-demand gap sequence for this observation period has been calculated. The value was 15 MW. Subsequently, the first-order difference sequence for that period was traversed, and the first-order difference values ​​corresponding to these five time indices were found to be 5 MW, 20 MW, 18 MW, -5 MW, and 16 MW, respectively. Based on the extraction logic, the data was compared one by one: The value of 5 MW is less than the standard deviation of 15 MW, so it is discarded; The value of 20 MW is greater than zero and greater than 15 MW, so it is retained; The value of 18 MW is greater than zero and greater than 15 MW, so it is retained; If the value is negative, discard it; The value of 16 MW is greater than zero and greater than 15 MW, so it is retained. Finally, it is precisely located to... These three target data points, and extract these three time indices to construct a set. .

[0034] The time indexes are spliced ​​together to generate continuous time segments, and these continuous time segments are stored sequentially into a list data structure to generate a time window set.

[0035] In the transformation stage of generating continuous analysis samples, a set containing discrete time indices is received. Since the flexible scheduling of adjustable resources typically requires maintaining a continuous time period to achieve hedging effects, isolated discrete data points are insufficient to support a complete evaluation loop. Therefore, a time adjacency judgment and span concatenation algorithm are required. First, all extracted time indices are sorted in ascending order of chronological order. Then, the time span difference between adjacent time indices is calculated. If the span difference between two adjacent time indices equals the base step size of a unified timestamp granularity, these two time indices are considered physically continuous. Boundary aggregation is performed on all continuous time indices to construct a closed time interval, i.e., a time continuous segment, using the start and end time indices of the continuous sequence. After span concatenation, a list data structure with linear order characteristics is constructed. The generated multiple time continuous segments are appended to this list data structure in tuple format according to their chronological order. This list data structure constitutes the final time window set in memory, serving as the sole computational view for subsequently constructing a resource-complementary directed graph.

[0036] For example, the output time index set Assume the basic step size for a unified timestamp granularity is 5 minutes. Step size validation is performed on this set, and it is found... minus This is exactly equal to one basic step size, indicating that the two are closely continuous in time; and minus If the distance is greater than one basic step size, it indicates a break between the two. Therefore, perform a span splicing operation: connect the continuous segments... and Aggregate to generate the first continuous time segment. ;because Isolated, it constitutes a second continuous time segment. Subsequently, the storage module initializes an empty list in memory and writes the two segments sequentially to form a list data structure. This list is the set of output time windows. It explicitly excludes valueless periods when the power grid is operating smoothly, greatly compresses the amount of correlation coefficient calculation for massive amounts of data, and directly projects it to the most critical supply and demand deterioration period.

[0037] Within the time window set, the correlation coefficient of the hourly output sequence is calculated. If the correlation coefficient is negative, its absolute value is used as the edge weight. A resource-complementary directed graph is constructed using the adjustable resources as nodes. The construction of the resource-complementary directed graph includes: Within the time window set, the covariance and standard deviation of the hourly output sequence are calculated, and the covariance is divided by the product of the two standard deviations to generate the Pearson correlation coefficient. In the calculation of the correlation between the operational behavior of adjustable resources, a time window set is extracted. This time window set defines the specific period when the grid regulation demand is most urgent. Within the time coordinate axis defined by this set, the hourly output sequences of any two adjustable resources are extracted in pairs, denoted as adjustable resource X and adjustable resource Y. To accurately measure the linear synchronous fluctuation characteristics of these two output sequences, the covariance and standard deviation of the hourly output sequences of adjustable resource X and adjustable resource Y are first calculated, and then the Pearson correlation coefficient is calculated. The calculation formula is as follows: , in, The Pearson correlation coefficient represents the relationship between adjustable resource X and adjustable resource Y. Its value range is strictly distributed between -1 and 1. Its physical meaning is to quantitatively characterize the degree of coordination or divergence in the output of the two adjustable resources during high-voltage operation of the power grid. The hourly output sequence representing adjustable resource X With the time-by-time output sequence of adjustable resource Y The covariance within the time window set comes from the statistical accumulation of the coordinated changing trends of the two sequences; The standard deviation of the hourly output sequence of resource X; The standard deviation represents the hourly output sequence of resource Y. Dividing the covariance by the scalar product of the two standard deviations eliminates the dimensional differences in the absolute order of installed capacity between different adjustable resources, reflecting waveform correlation.

[0038] For example, assume the time window set contains anomalous gap data for 20 time steps. Extract the hourly output sequences of adjustable resource A and adjustable resource B within these 20 time steps. After algebraic calculation, the covariance of the hourly output sequences of adjustable resource A and adjustable resource B is found to be -45. Simultaneously, the standard deviation of the hourly output sequence of adjustable resource A is calculated to be 8, and the standard deviation of the hourly output sequence of adjustable resource B is calculated to be 7.5. The calculation module uses the covariance as the dividend and the product of the two standard deviations as the divisor, performing scalar division to ultimately generate a Pearson correlation coefficient of -0.75 between adjustable resource A and adjustable resource B. Figure 2 As shown, the linear synchronous fluctuation pattern of different resources under extreme supply and demand conditions is mapped. The grid areas with coefficient values ​​less than zero in the figure represent resource pairs with significant complementary and hedging characteristics, providing a global data distribution view for extracting directed edge weights.

[0039] If the Pearson correlation coefficient is less than zero, perform absolute value transformation to generate edge weights, instantiate the adjustable resource as a graph structure node, connect the graph structure node with the edge weight as the connection attribute, and construct a resource complementary directed graph.

[0040] In the spatial topology generation stage, Boolean conditional filtering is performed based on the polarity of the calculated Pearson correlation coefficients. All paired Pearson correlation coefficients are iterated over, and if and only if the value of the Pearson correlation coefficient is less than zero, the corresponding two adjustable resources are determined to have a physically meaningful "complementary attribute," meaning that when the output of one resource decreases, exacerbating the grid burden, the output of the other resource can increase in the opposite direction to mitigate fluctuations. For Pearson correlation coefficients that satisfy this negative correlation condition (less than zero), their negative prefix is ​​removed, and an absolute value conversion operation is performed; the resulting positive scalar is the edge weight. Subsequently, each adjustable resource participating in the evaluation is instantiated as an independent data volume, i.e., a graph structure node, in memory space; and the previously generated edge weights are extracted and used as the connection attribute between two corresponding graph structure nodes. By traversing and connecting all graph structure nodes that meet the conditions, a complete resource complementarity directed graph is finally constructed at the data layer. In this graph structure, the closer the edge weight is to 1, the stronger the hedging and complementary ability of the two connected adjustable resources during periods of scarcity.

[0041] Based on the aforementioned resource-complementary directed graph, the betweenness centrality of each node is calculated, and hub weight coefficients are generated based on the betweenness centrality; the generation of hub weight coefficients includes: Based on the resource-complementary directed graph, calculate the total number of shortest paths between any two nodes, and count the frequency of the shortest path passing through the current node in the total number of shortest paths. Divide the frequency of the shortest path by the total number of shortest paths to generate betweenness centrality. In the network topology feature extraction stage, the intermediate transit status of each adjustable resource in the resource cooperation network is quantitatively evaluated based on the resource complementarity directed graph. To measure this topological bridging role, all node sets in the resource complementarity directed graph are traversed. For any node being evaluated in the network, any two source and target nodes in the network that do not contain the evaluated node are addressed, and the total number of shortest paths between the source and target nodes is calculated using breadth-first search or Dijkstra's algorithm. Subsequently, the connection attributes of these shortest paths are traversed and scanned, and the frequency of the shortest path passing through the current node is counted. The shortest path frequency is used as the dividend, and the total number of shortest paths is used as the divisor to perform scalar division. The division results of all source and target node pairings in the network are summed to finally generate the betweenness centrality of the current node, which is calculated using the following formula: , in, Representing the betweenness centrality of the currently evaluated node v, its physical meaning lies in quantifying the extent to which the adjustable resources corresponding to this node become the necessary path for other complementary resource pairs to effectively coordinate and hedge against each other. The higher the value, the stronger the irreplaceability of the adjustable resource during adjustment and scheduling; This represents the total number of shortest paths from source node s to target node t; This represents the frequency of the shortest path from source node s to target node t that passes through the currently evaluated node v. It maps the graph topology characteristics to specific numerical indicators based on objective network connectivity. For example, suppose a resource-complementary directed graph contains five graph nodes, each representing one of five independent adjustable resources. The betweenness centrality of the currently evaluated node c is calculated. Nodes a and e are extracted as the source and target nodes, respectively. The algorithm detects that there are a total of two shortest paths from node a to node e. Further tracing the node flow trajectories of these two paths reveals that one of the shortest paths passes through node c, meaning the frequency of the shortest path passing through node c is 1. Dividing the frequency 1 by the total number 2, we obtain the contribution value of the combination of nodes a and e to the centrality of node c as 0.5. The contributions of all similar node combinations in the resource-complementary directed graph to node c are summed. Assuming the final sum is 4.5, this 4.5 is confirmed as the betweenness centrality of node c.

[0042] The betweenness centrality is extracted to construct a centrality value sequence. The element with the largest value is extracted from the centrality value sequence to generate the maximum betweenness centrality value. The betweenness centrality is divided by the maximum betweenness centrality value to generate the hub weight coefficient.

[0043] In the feature normalization and weight allocation stage, the betweenness centrality generated by all graph structure nodes is read sequentially, and these values ​​are combined and loaded into a one-dimensional array according to the node's internal index order to construct a complete centrality value sequence. Since the absolute values ​​of the betweenness centrality calculated from resource-complementary directed graphs of different sizes vary considerably, to convert them into a unified dimensionless weight multiplier for subsequent multiplication in the option pricing model, an extreme value search operation is performed on the centrality value sequence to extract the element with the largest value, which is directly defined as the maximum betweenness centrality. Next, for each graph structure node, its own betweenness centrality is used as the dividend, and a division mapping operation is performed using the extracted maximum betweenness centrality as the common divisor. This operation directly generates the hub weight coefficient, and its core calculation formula is as follows: , in, The hub weight coefficient representing node v has its value range strictly standardized to a closed interval between zero and one, representing the relative monopoly premium ability of the adjustable resources corresponding to node v in the entire local complementary network. The betweenness centrality of node v; This represents the maximum betweenness centrality value dynamically extracted from the centrality value sequence. The most critical resource in the network's topology receives the highest weight, and the entire process requires no manually preset threshold parameters.

[0044] For example, in a complementary network scenario with six graph-structured nodes, the betweenness centrality of nodes A, B, C, D, and E are collected as 1.5, 3.0, 6.0, 1.2, 4.5, and 0.8, respectively. These five values ​​are concatenated sequentially to construct a centrality sequence [1.5, 3.0, 6.0, 1.2, 4.5, 0.8]. A maximum value retrieval algorithm is applied to this sequence, identifying the largest element as 6, and establishing it as the maximum betweenness centrality within the current time window. Subsequently, a final weight is assigned to each node: for node C, with a betweenness centrality of 6, dividing it by the maximum betweenness centrality of 6 yields a hub weight coefficient of 1, indicating that it is the absolute core mediator of this network area; for node B, with a betweenness centrality of 3.0, dividing it by 6 yields a hub weight coefficient of approximately 0.5, indicating that it has a moderately high hub status. Through pure data distribution-driven extreme value mapping, the value adjustment lever for each resource is objectively generated. Figure 3 As shown, the graph maps the intermediate transit status of different resources in a complementary topology network. The node values ​​at the top of the graph accurately reflect the pivotal support value of each resource for the coordinated regulation of the entire network, providing an intuitive data-driven basis for the weight allocation of the option pricing model.

[0045] Within the time window set, the maximum regulation capacity, maximum power ramp rate, and longest continuous regulation duration are extracted, divided by the global maximum value of the corresponding dimension across all nodes, and multiplied to generate a comprehensive regulation capability coefficient; the process includes: Within the time window set, the operating data of each node is extracted to parse out the maximum regulation capacity, maximum power ramp rate and longest continuous regulation duration, and the feature value sequence is constructed by category; In the adjustable resource physical performance feature extraction stage, a time window set is used as the time filter boundary, and the operating data of each node recorded by the underlying equipment within this boundary is retrieved. To comprehensively characterize the objective physical adjustment capability of the graph structure nodes, three core performance parameters are precisely parsed from the operating data. First is the maximum adjustment capacity, which represents the absolute power adjustment limit range that the resource can provide within a specific time period, typically measured in megawatts. Second is the maximum power ramp rate, which represents the resource's agility in responding to scheduling commands and changing its output, typically measured in megawatts per minute. Finally, the longest continuous adjustment duration, which represents the resource's endurance limit in maintaining maximum adjustable power under extreme power shortage conditions, typically measured in minutes or hours. After completing the parameter parsing of all nodes, the maximum adjustment capacity, maximum power ramp rate, and longest continuous adjustment duration extracted from each node are strictly categorized according to physical dimensions and horizontally concatenated into three independent one-dimensional arrays, thus constructing a feature value sequence corresponding to the three dimensions.

[0046] The process involves performing a maximum value retrieval operation on each of the feature value sequences to generate a global maximum value. Each feature value sequence is then divided by the global maximum value to generate a quotient value. Finally, the quotient values ​​are multiplied to generate a comprehensive adjustment capability coefficient. Specifically, the three generated feature value sequences are iterated over, and an extreme value addressing algorithm is used to extract the highest-valued term from each sequence. These three highest-valued terms are defined as the global maximum value for the corresponding dimension in the current evaluation batch. Subsequently, for each node, its three physical performance parameters are used as dividends, and scalar division is performed using the global maximum value for the corresponding dimension as the divisor, generating three quotient values ​​strictly between zero and one. This step completely eliminates the incommensurability between different physical dimensions. Finally, the three quotient values ​​belonging to the same node are multiplied together using a multiplication operator, and the resulting product is the comprehensive adjustment capability coefficient, calculated using the following formula: , in, The comprehensive adjustment capability coefficient generated by node v is a dimensionless index that reflects the overall lack of weaknesses in capacity, rate and endurance of the adjustable resources corresponding to the node. The closer the value is to 1, the better the overall adjustment quality. , , These represent the maximum regulation capacity, maximum power ramp rate, and longest continuous regulation duration, respectively, extracted from the operational data of node v. , , These represent the global maximum values ​​for the maximum regulation capacity, the maximum power ramp rate, and the longest sustained regulation duration, respectively, retrieved from the three feature value sequences. Through multiplicative aggregation logic, any resource with a severe deficiency in one aspect will have its overall regulation capacity coefficient penalized and lowered, ensuring that the final indicator is guided by the "weakest link" principle.

[0047] Within the time window set, the time mean of the first-order difference sequence is calculated. Within the time window set, all numerical items of the first-order difference sequence are extracted, and a summation operation is performed on all the numerical items to generate a total value. The total value is then divided by the time span of the time window set to generate the time mean.

[0048] In the macro-demand characteristic extraction stage, the corresponding first-order difference sequence is located and read from the time window set. To quantify the average adjustment urgency of the target power grid in the high-demand frequency band, all numerical terms contained in this first-order difference sequence are extracted. Subsequently, a scalar summation operation is performed on all numerical terms to generate a total sum. This total sum represents the cumulative amount of supply-demand gap deterioration within the entire time window set. Next, the time span contained in this time window set is obtained, which can be obtained by counting the total number of discrete time steps or the total duration. A scalar division operation is performed using the total sum as the dividend and the time span as the divisor to finally generate the time mean. The calculation formula is as follows: , in, The average value represents the time average. Its physical meaning lies in quantifying the average growth rate of the power grid's demand for regulation capacity during extreme periods, reflecting the intensity of demand in the external market environment. Represents the total value, which is derived from the absolute summation of all positive numerical terms in the first-order difference sequence within the time window set; This represents the time span, i.e., the total number of data points involved in the summation operation or the corresponding total time window step size. This method smooths discrete, dynamically differencing data into a globally static, scalar quantity.

[0049] For example, suppose the time window set generated after Boolean logic filtering contains five discrete five-minute time steps. Extract all numerical terms from the first-order difference sequence corresponding to these five time steps; the read values ​​are 10 MW, 15 MW, 20 MW, 18 MW, and 12 MW, respectively. Perform a summation operation on these five numerical terms sequentially, yielding a total of 75 MW. Since this interval contains five basic time steps, the time span is determined to be five. Subsequently, using 75 as the dividend and 5 as the divisor, the average generated time is calculated to be 15 MW per time step. This value is then used as a market demand multiplier that objectively reflects the urgency of the power grid.

[0050] The maximum regulating capacity, the time average, the hub weight coefficient, and the comprehensive regulating capacity coefficient are multiplied by scalars to generate option pricing feature values. The maximum regulating capacity, hub weight coefficient, and comprehensive regulating capacity coefficient are extracted and combined with the time average to construct a quaternary product formula. A scalar multiplication algebra operation is then performed to generate the option pricing feature values. For each node being evaluated in the network, four core input parameters are extracted: the first is the maximum regulating capacity, representing the upper limit of the absolute physical scale of the resource; the second is the hub weight coefficient, representing the irreplaceable intermediary role of the resource in the complementary graph topology; the third is the comprehensive regulating capacity coefficient, representing the multidimensional physical qualities of the resource in terms of capacity, rate, and endurance; and the fourth is the pre-calculated time average, representing the urgency of the objective demand imposed by the external power grid. These four feature parameters are constructed into a quaternary product formula, and a scalar multiplication algebra operation is directly performed. Since these four factors belong to four mutually orthogonal dimensions—physical scale, network topology, resource performance, and market demand—scalar multiplication algebraic operations can ensure that an exponentially amplified final evaluation value can be obtained if and only if the resource has no significant shortcomings in any of the four dimensions. The product of these multiplications generates the option pricing eigenvalue, calculated as follows: , in, It represents the option pricing feature value calculated for the current node, and is a highly aggregated relative scoring carrier used for resource selection and ranking. The maximum adjustment capacity of the representative node; Represents the average over time; The hub weight coefficient representing the node; This represents the comprehensive coefficient of a node's adjustment capability. Through the aforementioned quaternion multiplication mechanism, the traditional computing power model that relies solely on absolute capacity for evaluation is completely abandoned, achieving deep cross-domain integration driven by data.

[0051] For example, suppose we are currently evaluating a node in a distributed energy storage power station. We extract four parameters belonging to this node from the memory variable pool: the maximum regulation capacity resolved at the underlying level is 20 megawatts, the hub weight coefficient obtained from the prior graph theory algorithm is 0.8, the comprehensive regulation capacity coefficient obtained from physical performance normalization is 0.9, and the average time within the current time window set is 15. We substitute these four independent values ​​into a quaternion product formula to perform scalar multiplication algebraic operations, that is, perform... The multiplication instruction is used. After processing by the arithmetic unit, the result of the multiplication is 216. This value of 216 is the option pricing characteristic value of this energy storage power station node. This option pricing characteristic value will be fully saved, waiting to be used for horizontal comparison and arithmetic mean truncation and filtering with the characteristic values ​​of all other adjustable resource nodes in the area.

[0052] The effective regulation capacity assessment value of the target power grid is calculated by summing the maximum regulation capacity based on the option pricing characteristic values ​​of all nodes. The arithmetic mean of the option pricing characteristic values ​​is calculated for all nodes, and nodes whose option pricing characteristic values ​​are less than the arithmetic mean are removed. The maximum regulation capacity of the remaining nodes is summed, and the effective regulation capacity assessment value of the target power grid is output. Specifically, this includes: Extract the option pricing feature values, perform an accumulation operation to generate a feature sum, and divide the feature sum by the total number of nodes to generate an arithmetic mean; In the process of generating a dynamic evaluation benchmark, the option pricing feature values ​​calculated for each node in the graph structure are aggregated. To avoid introducing any subjective experience thresholds, a purely data-driven approach is used to calculate the overall value level of the current resource pool. First, all nodes participating in the evaluation are traversed in memory, their option pricing feature values ​​are extracted, and scalar summation is performed on these feature values. The accumulated result is the feature sum. Then, a counter is used to count the total number of nodes currently participating in the calculation. Using the feature sum as the dividend and the total number of nodes as the divisor, a scalar division operation is performed to generate the arithmetic mean, calculated using the following formula: , in, Representing the arithmetic mean, its physical meaning lies in quantifying the average comprehensive value level of all adjustable resources in the current target power grid area when dealing with extreme adjustment needs. This average value directly serves as the dynamic judgment threshold for subsequent implementation of survival of the fittest. The sum of representative features is derived from the absolute summation of the option pricing feature values ​​of all nodes; This represents the total number of nodes, i.e., the number of nodes in the current resource pool's graph structure. This averaging operation converges the multi-dimensional value scores into a single-dimensional macro-level benchmark.

[0053] For example, suppose that within a specific time window, the target power grid area has five adjustable resource nodes connected. The option pricing feature values ​​of these five nodes are extracted sequentially: 216, 150, 300, 50, and 100. These five values ​​are summed, yielding a total feature value of 816. Simultaneously, the total number of nodes is determined to be five through underlying counting. Then, 816 is divided by five, yielding an arithmetic mean of 163.2.

[0054] Traverse all nodes and compare the value of the option pricing feature with the arithmetic mean. Mark nodes whose option pricing feature is less than the arithmetic mean as invalid and remove them from the data. Extract the nodes that have not been removed and perform a scalar summation operation on the maximum regulating capacity. Output the effective regulating capacity assessment value of the target power grid.

[0055] In the screening and macro-capacity aggregation phase of high-quality regulating resources, a "suppression of false positives" operation is performed on the original resource pool based on the arithmetic mean. All nodes are traversed one by one, and the option pricing characteristic value of the currently being read node is compared with the arithmetic mean. If the comparison shows that the node's option pricing characteristic value is strictly less than the arithmetic mean, it indicates that the resource has significant shortcomings in physical performance, network complementary topology, or demand responsiveness, and its overall economic value is lower than the regional average. The status attributes of such nodes are marked as invalid in the underlying database, and data is removed from them, preventing them from participating in subsequent capacity statistics. Conversely, for nodes whose option pricing characteristic value is greater than or equal to the arithmetic mean, their status is determined as high-quality resources and they are retained, forming a set of retained nodes. Finally, for each retained node in this set, its maximum regulating capacity registered at the physical equipment level is extracted individually, and all extracted maximum regulating capacities are summed using a scalar operation. The total sum is the effective regulating capacity assessment value of the target power grid.

[0056] Example 2 like Figure 5 As shown, this embodiment of the invention provides an adjustable resource flexibility assessment device based on adjustable capacity, comprising: at least one processing unit, the processing unit being connected to a storage unit via a bus unit, the storage unit serving as a computer-readable storage medium, and capable of storing software programs, computer-executable programs, and modules, such as the software program, computer-executable program, and module corresponding to the adjustable resource flexibility assessment method based on adjustable capacity in this embodiment of the invention. The processing unit implements the aforementioned adjustable resource flexibility assessment method based on adjustable capacity by running the software program, computer-executable program, and module stored in the storage unit.

[0057] Of course, the computer program stored in the storage unit of the adjustable resource flexibility assessment device based on adjustable capacity provided in the embodiments of the present invention is not limited to the method operation described above, but can also execute related operations in the adjustable resource flexibility assessment method based on adjustable capacity provided in any embodiment of the present invention.

[0058] Example 3 Based on the same inventive concept, the present invention also provides an adjustable resource flexibility assessment system based on adjustable capacity, such as... Figure 4 As shown, the system includes: The time window generation module is used to obtain the hourly output sequence of adjustable resources and the hourly supply and demand gap sequence of the target power grid, calculate the first-order difference sequence and the standard deviation of the hourly supply and demand gap sequence, extract the continuous time segments in the first-order difference sequence that are positive and greater than the standard deviation of the hourly supply and demand gap sequence, and aggregate them to generate a time window set. The resource complementarity graph construction module is used to calculate the correlation coefficient of the hourly output sequence within the time window set. If the correlation coefficient is negative, the absolute value is taken as the edge weight, and a resource complementarity directed graph is constructed with the adjustable resources as nodes. The hub weight calculation module is used to calculate the betweenness centrality of each node based on the resource complementarity directed graph, and divide the betweenness centrality by the maximum betweenness centrality of the resource complementarity directed graph to generate hub weight coefficients. The regulation capacity calculation module is used to extract the maximum regulation capacity, the maximum power ramp rate and the longest continuous regulation duration within the time window set, divide them by the global maximum value of the corresponding dimension in all nodes and multiply them to generate a comprehensive regulation capacity coefficient. The option pricing feature value calculation module is used to calculate the time mean of the first difference sequence within the time window set, and multiply the maximum adjustment capacity, the time mean, the hub weight coefficient and the adjustment capability comprehensive coefficient by a scalar to generate the option pricing feature value. The effective regulation capacity assessment module is used to calculate the arithmetic mean of the option pricing characteristic values ​​of all nodes, remove nodes whose option pricing characteristic values ​​are less than the arithmetic mean, perform a summation operation on the maximum regulation capacity of the nodes that have not been removed, and output the effective regulation capacity assessment value of the target power grid.

[0059] In the embodiments provided by this invention, it should be understood that the disclosed structures and methods can be implemented in other ways. For example, the structural embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interfaces, structures, or units, and may be electrical, mechanical, or other forms.

[0060] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0061] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0062] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A method for assessing the flexibility of adjustable resources based on adjustable capacity, characterized in that, include: Obtain the hourly output sequence of adjustable resources and the hourly supply-demand gap sequence of the target power grid; Based on the first-order difference sequence and standard deviation of the hourly supply and demand gap sequence, a continuous time segment is extracted and aggregated to generate a time window set; Within the time window set, the correlation coefficient of the hourly output sequence is calculated. If the correlation coefficient is negative, the absolute value is taken as the edge weight, and a resource-complementary directed graph is constructed using the adjustable resources as nodes. The construction of the resource-complementary directed graph includes: within the time window set, calculating the covariance of the hourly output sequences of any two adjustable resources and the standard deviation of the hourly output sequences of any two adjustable resources; dividing the covariance by the product of the two standard deviations to generate the Pearson correlation coefficient; if the Pearson correlation coefficient is less than zero, it is determined that the corresponding two adjustable resources have complementary attributes; an absolute value transformation is performed to generate edge weights; the adjustable resources are instantiated as graph structure nodes; and the graph structure nodes are connected using the edge weights as connection attributes to construct the resource-complementary directed graph. Based on the resource-complementary directed graph, the betweenness centrality of each node is calculated, and hub weight coefficients are generated based on the betweenness centrality. This includes: for any node being evaluated in the network, addressing any two other source and target nodes in the network that do not contain the node being evaluated; calculating the total number of shortest paths between the source and target nodes; counting the frequency of the shortest path passing through the current node in the total number of shortest paths; dividing the frequency of the shortest path by the total number of shortest paths to generate the betweenness centrality of the node being evaluated; extracting the betweenness centrality to construct a centrality value sequence; extracting the element with the largest value from the centrality value sequence to generate the maximum betweenness centrality value; and dividing the betweenness centrality by the total number of shortest paths. The maximum value of the intermediate number centrality is used to generate the hub weight coefficient; within the time window set, the maximum regulation capacity, maximum power ramp rate, and longest continuous regulation duration are extracted, and the comprehensive regulation capability coefficient is calculated, including: within the time window set, extracting the operating data of each node to parse out the maximum regulation capacity, maximum power ramp rate, and longest continuous regulation duration, and aggregating them by category to construct a feature value sequence; performing a maximum value retrieval operation on each of the feature value sequences to generate a global maximum value, dividing each of the feature value sequences by the global maximum value to generate a quotient value, and performing a multiplication operation on the quotient value to generate the comprehensive regulation capability coefficient; within the time window set, calculating the time mean of the first-order difference sequence; Generating option pricing feature values ​​based on the maximum adjustment capacity, the time average, the hub weight coefficient, and the comprehensive adjustment capability coefficient includes: extracting all numerical terms of the first-order difference sequence within the time window set; performing a summation operation on all the numerical terms to generate a total value; dividing the total value by the time span of the time window set to generate a time average; extracting the maximum adjustment capacity, the hub weight coefficient, and the comprehensive adjustment capability coefficient; constructing a quaternary product formula with the time average; and performing a scalar multiplication algebra operation to generate option pricing feature values. The effective regulation capacity assessment value of the target power grid is calculated by summing the maximum regulation capacity based on the selection of valid nodes according to the option pricing feature value. This includes: extracting the option pricing feature value, performing an accumulation operation to generate a feature sum, dividing the feature sum by the total number of nodes to generate an arithmetic mean; traversing all nodes and comparing the value of the option pricing feature value with the arithmetic mean, marking nodes whose option pricing feature value is less than the arithmetic mean as invalid and removing them from the data, extracting the valid nodes that were not removed, extracting the maximum regulation capacity and performing a scalar summation operation, and outputting the effective regulation capacity assessment value of the target power grid.

2. The method for assessing the flexibility of adjustable resources based on adjustable capacity according to claim 1, characterized in that, The aggregated time window set includes: Raw metering data is collected for adjustable resources, and load deviation data is collected for the target power grid. Interpolation and alignment are performed at a uniform timestamp granularity to generate hourly output sequences and hourly supply-demand gap sequences. The root mean square deviation of the hourly supply and demand gap sequence is calculated as the standard deviation of the hourly supply and demand gap sequence; A first-order difference sequence is generated by subtracting the preceding term from the following term in the hourly supply-demand gap sequence. Based on the first-order difference sequence and standard deviation of the hourly supply and demand gap sequence, a continuous time segment is extracted and aggregated to generate a time window set.

3. The method for assessing the flexibility of adjustable resources based on adjustable capacity according to claim 2, characterized in that, The aggregated time window set includes: Extract target data points from the first-order difference sequence that have positive values ​​and are greater than the standard deviation, and extract the time index of the target data points; The time indices are concatenated to generate continuous time segments. These continuous time segments are then stored sequentially in a list data structure to generate a time window set. All extracted time indices are sorted in ascending order by time. The time span difference between adjacent time indices is calculated. If the span difference between two adjacent time indices equals the base step size of the unified timestamp granularity, these two time indices are considered to be physically continuous. Boundary aggregation is performed on all continuous time indices, and a closed time interval is constructed using the start and end time indices of the continuous sequence as the continuous time segment.

4. The method for assessing the flexibility of adjustable resources based on adjustable capacity according to claim 1, characterized in that, The total number of shortest paths between the source node and the target node can be calculated using breadth-first search or Dijkstra's algorithm.

5. A system for assessing the flexibility of adjustable resources based on adjustable capacity, characterized in that, include: The time window generation module is used to obtain the hourly output sequence of adjustable resources and the hourly supply and demand gap sequence of the target power grid. Based on the first-order difference sequence and standard deviation of the hourly supply and demand gap sequence, the continuous time segment is extracted and aggregated to generate a time window set. The resource complementarity graph construction module is used to calculate the correlation coefficient of the hourly output sequence within the time window set. If the correlation coefficient is negative, the absolute value is taken as the edge weight, and a resource complementarity directed graph is constructed with the adjustable resources as nodes. This includes: within the time window set, calculating the covariance of the hourly output sequences of any two adjustable resources and the standard deviation of the hourly output sequences of any two adjustable resources; dividing the covariance by the product of the two standard deviations to generate the Pearson correlation coefficient; if the Pearson correlation coefficient is less than zero, determining that the corresponding two adjustable resources have complementary attributes; performing absolute value transformation to generate edge weights; instantiating the adjustable resources as graph structure nodes; and connecting the graph structure nodes with the edge weights as connection attributes to construct a resource complementarity directed graph. The hub weight calculation module is used to calculate the betweenness centrality of each node based on the resource complementarity directed graph, and generate hub weight coefficients based on the betweenness centrality. This includes: based on the resource complementarity directed graph, for any node to be evaluated in the network, addressing any two other source nodes and target nodes in the network that do not contain the node to be evaluated; calculating the total number of shortest paths between the source nodes and the target nodes; counting the frequency of the shortest path passing through the current node in the total number of shortest paths; dividing the frequency of the shortest path by the total number of shortest paths to generate the betweenness centrality of the node to be evaluated; extracting the betweenness centrality to construct a centrality value sequence; extracting the element with the largest value from the centrality value sequence to generate the maximum betweenness centrality value; and dividing the betweenness centrality by the maximum betweenness centrality value to generate the hub weight coefficients. The regulation capacity calculation module is used to extract the maximum regulation capacity, maximum power ramp rate, and longest continuous regulation duration within the time window set, and calculate the comprehensive regulation capacity coefficient. This includes: extracting the operating data of each node within the time window set to parse out the maximum regulation capacity, maximum power ramp rate, and longest continuous regulation duration; aggregating these data by category to construct a feature value sequence; performing a maximum value retrieval operation on each feature value sequence to generate a global maximum value; dividing each feature value sequence by the global maximum value to generate a quotient; and performing a multiplication operation on the quotient to generate the comprehensive regulation capacity coefficient. The option pricing feature value calculation module is used to calculate the time mean of the first-order difference sequence within the time window set, and generate option pricing feature values ​​based on the maximum adjustment capacity, the time mean, the hub weight coefficient, and the comprehensive adjustment capability coefficient. The module includes: extracting all numerical terms of the first-order difference sequence within the time window set; performing a summation operation on all numerical terms to generate a total value; dividing the total value by the time span of the time window set to generate the time mean; extracting the maximum adjustment capacity, the hub weight coefficient, and the comprehensive adjustment capability coefficient; constructing a quaternary product formula with the time mean; and performing a scalar multiplication algebra operation to generate the option pricing feature value. The effective regulation capacity assessment module is used to filter valid nodes based on option pricing feature values ​​to summarize the maximum regulation capacity and calculate the effective regulation capacity assessment value of the target power grid. This includes: extracting the option pricing feature values, performing an accumulation operation to generate a feature sum, dividing the feature sum by the total number of nodes to generate an arithmetic mean; traversing all nodes and comparing the option pricing feature values ​​with the arithmetic mean, marking nodes whose option pricing feature values ​​are less than the arithmetic mean as invalid and removing them from the data, extracting the valid nodes that were not removed, extracting the maximum regulation capacity and performing a scalar summation operation, and outputting the effective regulation capacity assessment value of the target power grid.

Images