Frequency modulation resource partition allocation method and system considering frequency distribution

Abstract

The invention relates to the technical field of power system safety control, in particular to a frequency modulation resource partition allocation method and system considering frequency distribution, and the method comprises the steps: obtaining a frequency dynamic index of each power system node, and calculating the maximum frequency difference of the power system nodes; dynamically dividing a fault area and a non-fault area according to the position of the disturbance source; establishing an optimization function with the goal of minimizing the maximum value in the maximum frequency difference of the nodes of the power system; on the basis of the optimization function, according to the maximum frequency difference of the nodes of the power system, inertia distribution of the power system, network topology and electrical distance parameters, fast frequency response resources with different capacities are distributed to the fault area and the non-fault area through a multi-factor constraint optimization model; and configuring a fast frequency response resource of which the density is higher than a preset threshold value in the fault area so as to realize localized suppression of the frequency deviation. According to the invention, the problem of over-matching or under-matching of resources caused by static partition and a fixed threshold value distribution mechanism in the prior art is effectively solved.

Description

Technical Field

[0001] This invention relates to the field of power system safety control technology, and in particular to a method and system for partitioning and allocating frequency regulation resources considering frequency distribution. Background Technology

[0002] In the field of power system frequency security control, when the system encounters power disturbances, the frequency dynamic response characteristics are strongly correlated with the electrical location, inertia distribution, and network topology of each node. Therefore, it is necessary to suppress the spread of frequency deviation by configuring fast frequency response resources (such as frequency regulating units and energy storage systems). Existing technologies mainly use allocation mechanisms based on fixed partitioning patterns, that is, pre-dividing static areas according to the geographical or administrative boundaries of the power network, and uniformly configuring resources for all nodes in the preset area according to fixed capacity thresholds after a fault occurs; and global uniform allocation strategies, ignoring the dynamic characteristics of the system, and uniformly deploying frequency regulating resource capacity throughout the entire network.

[0003] However, the aforementioned existing technologies suffer from several drawbacks. Rigid partitioning leads to a disconnect between resources and fault propagation characteristics. Static partitioning fails to consider the dynamic changes in the electrical location of the disturbance source and its propagation path, making it difficult to match resource allocation to the actual fault impact range. This results in insufficient resource supply at critical weak points rather than capacity redundancy in critical areas. Furthermore, the threshold allocation mechanism lacks dynamic coupling analysis; fixed capacity thresholds do not incorporate the attenuation effect of network topology connectivity and electrical distance on frequency deviation transmission, leading to inaccurate resource coverage of highly sensitive disturbance propagation paths. Global homogenization exacerbates system imbalance; the unified allocation strategy ignores the priority of inertia support needs in local fault areas, preventing limited frequency regulation resources from effectively blocking the chain propagation of frequency deviations. These shortcomings collectively contribute to a decrease in the transient stability of the power system under disturbances, especially in large-scale renewable energy integration scenarios, where suboptimal frequency response resource allocation methods significantly increase system recovery costs.

[0004] The information disclosed in this background section is intended only to enhance the understanding of the general background of this disclosure and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0005] This invention provides a frequency modulation resource partitioning allocation method and system that takes into account frequency distribution, which can effectively solve the problems in the background art.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A method for partitioning and allocating frequency modulation resources considering frequency distribution, the method comprising: Obtain the frequency dynamic indicators of each power system node in the power system and calculate the maximum frequency difference of the power system node; Based on the location of the disturbance source, the power system is dynamically divided into fault areas and non-fault areas, wherein the fault area includes the disturbance source node and a set of nodes within a preset electrical distance threshold range; Establish an optimization function with the objective of minimizing the maximum value among the maximum frequency differences of the nodes in the power system; Based on the optimization function, and according to the maximum frequency difference of the power system nodes, the power system inertia distribution, network topology, and electrical distance parameters, a multi-factor constraint optimization model is used to allocate fast frequency response resources of different capacities to the fault region and the non-fault region. In the fault region, fast frequency response resources with a density higher than a preset threshold are configured to achieve localized suppression of frequency deviation.

[0007] Furthermore, obtain the frequency dynamic indicators of each power system node in the power system, including: Based on real-time measurement data from synchronous phasor measurement, the maximum frequency change rate and frequency drop depth of the power system nodes are obtained as the frequency dynamic indicators during time-domain simulation. The transient bias of the time-domain simulation is corrected based on the correlation between the power angle trajectory and frequency response of the power system nodes. The maximum frequency difference of the power system node is calculated based on the corrected maximum rate of frequency change and the frequency drop depth.

[0008] Furthermore, the power system is dynamically divided into fault regions and non-fault regions based on the location of the disturbance source, including: Calculate the distance relationship metric between all the power system nodes and the disturbance source nodes based on the electrical distance parameters; Based on the comparison between the distance relationship metric and the preset electrical distance threshold, nodes whose electrical coupling strength meets the preset requirements are selected to form an initial candidate set; Based on the adjacency matrix of the network topology, verify the reachability of each node in the initial candidate set to the disturbance source node according to the topological path, and exclude nodes that do not meet the connectivity constraints. The verified nodes are integrated with the disturbance source nodes to form a complete set of nodes for the fault region.

[0009] Furthermore, configuring high-frequency response resources with a density higher than a preset threshold in the fault region includes: Obtain the capacity allocation indices of the fault region and the non-fault region output by the multi-factor constrained optimization model; Calculate the unit node resource capacity of the faulty region and the non-faulty region; Based on the distribution density of key topological location nodes in the complete node set of the fault region, adjust the resource capacity of the unit node in the fault region. Based on the comparison between the adjusted unit node resource capacity of the faulty region and the unit node resource capacity of the non-faulty region, a resource allocation scheme that satisfies the density exceeding the preset threshold is confirmed, and the final configuration instruction is output based on the localized suppression of the fast frequency response resource.

[0010] Furthermore, based on a multi-factor constrained optimization model, different capacities of fast frequency response resources are allocated to the faulty region and the non-faulty region, including: Based on the convergence objective of the optimization function, the complete set of nodes in the fault region is loaded as a spatial domain boundary constraint into the multi-factor constraint optimization model. Based on the correspondence between the identification results of inertial weak points in the complete node set of the fault region and the inertia distribution of the power system, a priority factor for resource allocation in the fault region is generated. Based on the path connectivity characteristics of the network topology, verify the impact of the priority factor distribution on the transmission attenuation effect in the non-faulty area; Based on the optimization function, the resource allocation instructions generated by the multi-factor constraint optimization model in response to the priority factor and the influence of transmission attenuation are configured to allocate fast frequency response resource capacity with a dominant optimization level to the fault region, and supplementary capacity resources that conform to the dominant optimization level are configured to the non-fault region.

[0011] Furthermore, configuring high-frequency response resource capacity with a dominant optimization level to the fault region includes: Spatial mapping association is performed between the complete set of nodes in the fault region and the fast frequency response resource capacity of the dominance optimization level. Based on the transitional connection attributes of nodes in the network topology on the energy transmission path, identify the topological connection attribute nodes in the complete node set of the fault area. For the nodes with topological connection attributes, topological fluid equilibrium weights are added to the priority factor according to the spatial association algorithm of the multi-factor constraint optimization model; The fast frequency response resource capacity of the dominant optimization level is reallocated according to the priority factor after the weighting, so that the topology connection attribute node obtains a proportion-increased configuration capacity share. The configured capacity share is written into the execution control parameters of the localized suppression of the frequency deviation.

[0012] Furthermore, the transient bias of the time-domain simulation is corrected based on the constructed correlation between the power angle trajectory and frequency response of the power system nodes, including: Identify the transient power angle swing range of each power system node in the time-domain simulation and obtain the phase amplitude related variables of the power angle trajectory; The phase amplitude-related variables of each power system node are mapped to the oscillating energy transfer function established for the frequency response, thereby generating the angular frequency coupling strength index. The non-monotonic distortion component of the frequency drop depth is corrected based on the power angle frequency coupling strength index. The frequency drop depth after distortion correction is dynamically weighted by the power angle frequency coupling strength index to generate the compensated weighted frequency drop depth. The maximum frequency difference is calculated by using the compensated weighted frequency drop depth and the maximum value of the corrected frequency change rate.

[0013] Furthermore, adjusting the resource capacity of the unit node in the faulty region includes: Identify the topological location attributes of each node in the complete node set of the fault area, and output the location configuration coordinates of the key topological location nodes; Calculate the connectivity distribution index of the key topological location nodes based on the location configuration coordinates; The resource capacity allocation of the unit node is adjusted using a gradient calculation based on the connectivity distribution index. Verify whether the unit node resource capacity after gradient adjustment operation enhances the node distribution density advantage of the fault region; The aggregated effect of the enhanced unit node resource capacity on the complete set of nodes in the fault region is used as the adjusted unit node resource capacity input into the resource allocation scheme.

[0014] A frequency modulation resource partitioning allocation system considering frequency distribution, the system comprising: The information acquisition module acquires the frequency dynamic indicators of each power system node in the power system and calculates the maximum frequency difference of the power system nodes; The region division module dynamically divides the power system into fault regions and non-fault regions based on the location of the disturbance source. The fault region includes the disturbance source node and a set of nodes within a preset electrical distance threshold. The function optimization module establishes an optimization function with the objective of minimizing the maximum value among the maximum frequency differences of power system nodes; The resource allocation module, based on an optimization function, allocates fast frequency response resources of different capacities to faulty and non-faulty areas through a multi-factor constrained optimization model according to the maximum frequency difference of power system nodes, power system inertia distribution, network topology, and electrical distance parameters. The local suppression module configures fast frequency response resources with a density higher than a preset threshold in the fault area to achieve localized suppression of frequency deviation.

[0015] Furthermore, the information acquisition module includes: The time-domain simulation unit, based on real-time measurement data from synchronous phasor measurements, obtains the maximum frequency change rate and frequency drop depth of power system nodes as dynamic frequency indicators during the time-domain simulation process. The deviation correction unit corrects the transient deviation of the time-domain simulation based on the correlation between the power angle trajectory and frequency response of the constructed power system nodes. The frequency difference calculation unit calculates the maximum frequency difference of power system nodes based on the corrected maximum frequency change rate and frequency drop depth.

[0016] The technical solution of this invention can achieve the following technical effects: By establishing an optimization function aimed at minimizing the maximum frequency difference of system nodes, the fault regions are dynamically divided by integrating electrical distance and network topology characteristics. The multi-factor constrained optimization model is then driven to configure fast frequency response resources with dominant optimization levels in the fault regions. This solves the problem of resource over-allocation or under-allocation caused by static partitioning and fixed threshold allocation mechanisms in existing technologies, and achieves precise local suppression of frequency deviation.

[0017] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0018] 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, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating a frequency modulation resource partitioning allocation method that takes frequency distribution into account. Figure 2 A flowchart illustrating the process of obtaining frequency dynamic indicators; Figure 3A flowchart illustrating the division of faulty and non-faulty areas; Figure 4 A flowchart illustrating the process of correcting transient biases in time-domain simulation. Detailed Implementation

[0020] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0022] Example 1; like Figure 1 As shown, this application provides a frequency modulation resource partitioning allocation method considering frequency distribution, the method including: S10: Obtain the frequency dynamic indicators of each power system node in the power system and calculate the maximum frequency difference of the power system node; S20: Based on the location of the disturbance source, the power system is dynamically divided into fault areas and non-fault areas, wherein the fault area includes the disturbance source node and the set of nodes within the preset electrical distance threshold range; S30: Establish an optimization function with the objective of minimizing the maximum value among the maximum frequency differences of the nodes in the power system; S40: Based on the optimization function, according to the maximum frequency difference of power system nodes, power system inertia distribution, network topology and electrical distance parameters, a multi-factor constrained optimization model is used to allocate fast frequency response resources of different capacities to fault areas and non-fault areas. S50: Configure fast frequency response resources with a density higher than a preset threshold in the fault area to achieve localized suppression of frequency deviation.

[0023] Specifically, firstly, high-precision synchronous phasor measurement devices are deployed in the power system, and combined with conventional measurement data, real-time frequency dynamic indicators of each node are acquired, such as instantaneous frequency, frequency change rate, and frequency valley. By performing window function processing and filtering on the time-series data after a disturbance occurs, the maximum frequency difference of each node under the disturbance condition is determined, i.e., the maximum absolute frequency deviation during the disturbance process. In the preferred implementation, high sampling rate data from the synchronous phasor measurement devices is used to improve the resolution of the frequency change rate, and the measured indicators are corrected using short-time simulations or historical disturbance databases to obtain a more reliable estimate of the maximum frequency difference. Subsequently, based on the location of the disturbance source... The electrical distance between nodes is calculated based on the power grid topology and electrical parameters. Equivalent migration impedance or sensitivity-based impedance distance measurement can be used. A preset electrical distance threshold is used to determine fault and non-fault areas. The preferred threshold can be an empirical value or adaptively adjusted based on the disturbance magnitude. For example, a larger electrical distance threshold is used for larger disturbances to include more affected nodes, while a smaller threshold is chosen for smaller disturbances to limit the response range. In practical engineering, the initial value of this threshold can be determined through prior operating conditions or online sensitivity analysis, and then gradually corrected based on observation data during operation. For the optimization objective, a system is established to minimize... The maximum value of the maximum frequency deviation of all nodes in the network is the objective optimization function. This means that reducing the maximum frequency deviation of the worst-performing node in the system is the primary objective. Weights are introduced into the objective function to account for performance indicators such as frequency valleys, frequency change rate, and recovery time. For ease of engineering implementation, a sequential optimization strategy is preferred in the solver: first, reducing the maximum frequency deviation is the primary objective, followed by reducing the weighted average frequency deviation of the entire network and decreasing the frequency recovery time as secondary objectives. Regarding constraint settings, the maximum output, duration, and response delay of each fast frequency response resource, resource availability, grid safety constraints, and economic considerations are comprehensively taken into account. Furthermore, the inertia distribution of the power system, network topology, and electrical distance parameters are integrated into the constraint system as influencing factors to ensure the feasibility and robustness of the allocation scheme. For the solution method, a hybrid optimization strategy of heuristic and deterministic optimization under multi-factor constraints is preferred: in the initial stage, heuristic algorithms, such as genetic algorithms or particle swarm optimization, can be used to quickly explore feasible allocation solutions in the global space to avoid getting trapped in local optima. Subsequently, deterministic optimization methods, such as sequential quadratic programming or feasible direction method, are used to locally refine the heuristic results, and finally output the fast frequency response resource capacity allocation schemes for fault areas and non-fault areas respectively.To improve online response capabilities, this optimization process can be combined with a fast unpacking library of several typical disturbance conditions generated offline in advance. This allows for the rapid invocation of the allocation scheme with the closest similarity to the disturbance when it occurs, and minor online adjustments can be made. To achieve localized suppression of frequency deviation, this implementation requires that the density of fast frequency response resources configured in the fault area be higher than a preset threshold. This density threshold can be determined based on the average inertia of the nodes, the influence radius of the fault node, and the total available frequency regulation resources of the system. For example, in a transmission network example, when a fault occurs at an important node and the equivalent inertia of the nodes within three hops around that node is low, a larger proportion of the total available fast response capacity can be allocated to the fault area to ensure that the resource density per unit node or unit line length in the area exceeds the preset threshold. This allows local resources to mainly undertake the recovery task during the frequency trough phase, reducing the large-scale use of remote resources and shortening the frequency recovery time. To verify the effectiveness of the solution, scenario testing at the dispatch center is recommended: using typical disturbances, such as three-phase fault clearing of units or sudden line tripping, as the object, first use simulation models with different inertia distributions and response configurations to evaluate the impact of the generated partition configuration on the maximum frequency difference and frequency recovery curve. Based on this, continuously adjust the electrical distance threshold, fault area resource density threshold, and optimization weights until a small worst-case node frequency difference can be achieved under various disturbance conditions, satisfying operational and economic constraints.

[0024] The technical solution of this invention establishes an optimization function with the goal of minimizing the maximum frequency difference of system nodes, dynamically integrates electrical distance and network topology characteristics to divide fault areas, and drives a multi-factor constraint optimization model to configure fast frequency response resources with dominant optimization level in the fault areas. This solves the problem of resource over-allocation or under-allocation caused by static partitioning and fixed threshold allocation mechanisms in existing technologies, and achieves accurate local suppression of frequency deviation.

[0025] Furthermore, such as Figure 2 As shown, the frequency dynamic indicators of each power system node in the power system are obtained, including: Based on real-time measurement data from synchronous phasor measurement, the maximum frequency change rate and frequency drop depth of power system nodes are obtained as dynamic frequency indicators during time-domain simulation. The transient bias of the time-domain simulation is corrected based on the correlation between the power angle trajectory and frequency response of the power system nodes. The maximum frequency difference of power system nodes is calculated based on the corrected maximum rate of frequency change and the frequency drop depth.

[0026] As a preferred embodiment of the above, firstly, synchronous phasor measurement devices are deployed at important nodes of the power system, and their time synchronization accuracy is ensured. This allows for the acquisition of high-sampling-rate time-series measurement data from all measurement points across the entire network when disturbances occur. The preferred sampling rate is tens to hundreds of samples per second to meet the resolution requirements for frequency change rates. The raw phasor data is then preprocessed, including narrowband filtering for noise reduction, interpolation of missing data, removal of abrupt outliers, and precise alignment of all measurement point data based on timestamps. After preprocessing, a pre-established time-domain simulation model is used to simulate the disturbance scenario to be identified. The simulation employs the same event start time and initial operating conditions as the measurement system. The simulation output is used in parallel with the measured frequency curves to acquire the data of each power system node. Using peak frequency change rate and frequency drop depth as initial frequency dynamic indicators, this implementation proposes a correction method based on constructing the correlation between power angle trajectory and frequency response to address potential transient deviations in time-domain simulations. On one hand, the power angle time-series trajectory of each generator or node is reconstructed from the measured voltage phase data through phase relationship conversion with a reference node or generator terminal, or through dynamic state estimation methods with a network model. On the other hand, the power angle trajectory of the corresponding node or unit is extracted from the simulation results, and the main types of simulation transient deviations, such as amplitude scaling, time delay, or spectral distortion, are identified using the relative amplitude difference, phase shift, and inconsistency in the peak occurrence time as indicators. Subsequently, the simulation is corrected through guided parameter adjustments, preferably employing... By adjusting identifiable parameters such as equivalent inertia, damping, speed governor, primary frequency regulation gain, and time constant, the simulated power angle trajectory is made consistent with the measured trajectory within a specified error tolerance. Parameter adjustment can be achieved using inverse identification based on least squares or iterative optimization based on gradients. After each iteration, the root mean square error and peak deviation of the frequency curve are compared until the error is below a preset threshold. When field measurement coverage is incomplete, this implementation method preferably uses interpolation or dynamic state estimation methods based on network topology and equivalent generator parameters to estimate the power angle and frequency response of unmeasured nodes, ensuring that all nodes can obtain correction data. After power angle trajectory correction is completed, the maximum frequency change rate and frequency drop of each node are extracted from the corrected time series. The maximum frequency difference of a node is calculated based on the correction results of the two factors. The calculation method is to determine the peak value of the absolute frequency offset throughout the entire disturbance process and perform window function smoothing on possible short-term fast pulses to avoid misjudgment of instantaneous noise. For engineering implementation, it is preferable to perform multiple disturbance sensitivity simulations on the same disturbance within the parameter uncertainty range to obtain a set of corrected frequency indices and take the conservative value of them as the maximum frequency difference of the node to ensure that the frequency difference estimation under the most unfavorable situation is adopted when allocating resources. For ease of application, this embodiment also provides several preferred engineering parameters and criteria as examples: the sampling rate of the synchronous phasor measurement device is preferably fifty to two hundred times per second, the filter adopts bandpass or low-pass to retain short-time dynamic characteristics, and the filter window is preferably 0.2 to 1 second; the error tolerance for comparing the power angle trajectory with the simulation can be set to a root mean square error of frequency not exceeding 0.01 Hz or a root mean square error of power angle path not exceeding 0.01 radians; the termination condition for parameter identification is that the error decrease is lower than the minimum threshold or the maximum number of iterations is reached in several consecutive iterations; for example: when a unit disconnection event occurs in a typical 150 bus system, the field PMU measures that a key bus experiences a maximum frequency drop of 0.48 Hz and a maximum frequency change rate of 0.35 Hz per second in the early stage of the disturbance. Simulation comparison reveals that the simulated frequency valley is too small and accompanied by an earlier simulated power angle peak of approximately 0.1 seconds. After adjusting the equivalent inertia and primary frequency modulation gain in the simulation and iterative identification, the simulated and measured power angle trajectories achieve good overlap. After correction, the maximum frequency difference of this node is determined to be 0.52 Hz; this value is then used as the input parameter in subsequent partition allocation optimization.

[0027] Furthermore, such as Figure 3 As shown, the power system is dynamically divided into fault regions and non-fault regions based on the location of the disturbance source, including: Calculate the distance relationship between all power system nodes and disturbance source nodes based on electrical distance parameters; Based on the comparison between the distance relationship metric and the preset electrical distance threshold, nodes whose electrical coupling strength meets the preset requirements are selected to form an initial candidate set. Based on the adjacency matrix of the network topology, the reachability of each node in the initial candidate set to the disturbance source node according to the topological path is verified, and nodes that do not meet the connectivity constraints are excluded. The verified nodes are integrated with the disturbance source nodes to form a complete set of nodes in the fault region.

[0028] As a preferred embodiment of the above, firstly, electrical distance parameters are constructed based on system state data and network parameters, and distance relationship metrics from all nodes to disturbance source nodes are calculated. Preferably, equivalent line impedance or coupling strength based on power flow sensitivity is used as the basis for electrical distance measurement. Simultaneously, line impedance and power flow distribution information can be weighted to synthesize a composite distance, taking into account both transmission physical distance and power flow coupling characteristics. In engineering implementation, the original network parameters are preprocessed to remove temporarily disconnected branches, and sparse matrix storage is used to improve computational efficiency. Secondly, the distance relationship metrics are compared with a preset electrical distance threshold, and an initial candidate set is selected. The preferred threshold can be an absolute value, or it can be a quantile or ranking method. To determine the optimal approach, nodes with the smallest distance metrics, such as the top percentages or top-ranked nodes, can be selected as candidates. Alternatively, nodes with equivalent impedance less than a certain threshold can be selected. To improve robustness, an adaptive strategy related to the magnitude of the disturbance is preferred in the threshold setting. The larger the disturbance, the wider the threshold is to include a broader range of influence; the smaller the disturbance, the tighter the threshold to achieve a localized response. The third step uses the adjacency relationships of the network topology to verify the topological reachability of the initial candidate set one by one. Preferably, a directed or undirected graph is constructed using the adjacency matrix of the current network state, and graph algorithms such as shortest path search or breadth-first traversal are used to verify whether each candidate node is reachable from the disturbance source node on the topological path. During verification, line connectivity and primary / backup configurations are considered simultaneously. Using line status and the connectivity rules of converter stations or substations, if a node is unreachable under the current operating conditions due to intermediate line interruption or network separation, it is excluded from the candidate set. During the verification process, a weighted adjacency matrix is ​​preferably used, taking line capacity or commutation limitations as edge weights to avoid including nodes in the fault region through capacity-limited paths. Finally, nodes that have passed electrical coupling strength screening and topology reachability verification are integrated with disturbance source nodes to form a complete node set for the fault region. Simultaneously, a secondary judgment is preferably performed on boundary nodes: nodes that are close to the distance threshold and topologically reachable but have low frequency response sensitivity can be temporarily included in secondary candidates, and their inclusion in the fault region will be determined in subsequent dynamic sensitivity analysis or PMU collaborative observation. To address situations where measurements are missing or topology information is delayed, conservative approaches can be taken to handle unresolved nodes to ensure safety, using either typical influence radii based on historical operating conditions or alternative strategies based on neighborhood clustering. For example, in a simulation test using a high-voltage bus as the disturbance source, equivalent impedance was used as the electrical distance metric, and an initial threshold was set so that the top 20% of nodes by distance would enter the candidate set, resulting in 40 candidate nodes. Subsequently, using the current network topology adjacency matrix and considering the actual operating condition that one of the two parallel lines had tripped, five nodes isolated due to connectivity were eliminated through shortest path verification, ultimately generating a fault area set containing 35 nodes. This set was then used for subsequent localized frequency regulation resource allocation.

[0029] Furthermore, configuring high-frequency response resources with a density exceeding a preset threshold in the fault area includes: Obtain the capacity allocation index of the fault region and non-fault region output by the multi-factor constrained optimization model; Calculate the resource capacity of a unit node in both faulty and non-faulty regions; Based on the distribution density of key topological locations in the complete node set of the fault region, adjust the resource capacity per node in the fault region. Based on the comparison between the unit node resource capacity of the adjusted faulty region and the unit node resource capacity of the non-faulty region, a resource allocation scheme that satisfies the density exceeding the preset threshold is confirmed, and the final configuration command is output based on the localized suppression of fast frequency response resources.

[0030] As a preferred embodiment of the above, the output of the multi-factor constrained optimization model is first received and parsed. This output includes information such as the overall capacity allocation index for faulty and non-faulty regions, the types and available output of various response resources, response delay, and duration. Based on this output, the unit node resource capacity for faulty and non-faulty regions is preferably calculated using a node-weighted method. Specifically, the total allocated fast response capacity within the region is divided by the number of nodes in that region, and a weight correction is applied to the importance of nodes during the calculation. The weight can be determined by the node's equivalent inertia, load importance, or the concentration of generators or converter stations to reflect the different contributions of different nodes to frequency stability. Subsequently, based on... The distribution density of key topological locations in the fault area node set is used to adjust the resource capacity of the above-mentioned unit nodes: preferably, key topological location nodes are identified by graph theory indicators, such as nodes with high degree centrality, high betweenness centrality, nodes connecting multiple transmission channels or adjacent to high-power loads or important power generation units, and the distribution density of these key nodes in the fault area is calculated. If the key nodes are concentrated and sparsely distributed, the unit node capacity of the cell where these nodes are located is increased first. The adjustment strategy can adopt a hierarchical weighted approach, that is, an additional capacity bonus is allocated to the location of key nodes to ensure local response capability, and the adjustment is simultaneously constrained by actual constraints such as resource availability, maximum output of individual units and response rate. To ensure the adjustment results meet the requirement of density exceeding a preset threshold, the judgment rule is to compare the unit node resource capacity of the adjusted faulty area with the corresponding value of the non-faulty area, and then compare it with the preset density threshold. This threshold can be an absolute capacity, a node value, or the minimum allowable value of the capacity ratio between the faulty area and the non-faulty area. If the threshold is still not met after adjustment, remedial measures are taken in order of priority, including reallocating movable or redistributable rapid response resources within the faulty area, activating short-term sharing protocols for nearby nodes, or issuing additional energy storage or controllable load response commands from the scheduling layer. In extreme cases of severe resource shortage, it is preferable to call upon a pre-established external emergency resource pool or request manual intervention to ensure safety. After completing capacity adjustment and threshold verification at the boundary, the final configuration command is output based on the type and control characteristics of the fast frequency response resource. The command includes the target output, response curve priority, triggering conditions and duration of each response unit. It is preferred to use the current communication and control standards of the power system for issuance and execution, such as using the dispatch automation system to send settings to the energy storage controller, adjust the fast frequency response parameters of the renewable energy inverter, or issue load instantaneous trip or demand-side response actions. At the same time, before issuance, the configuration scheme is verified online in the time domain based on the modified frequency dynamic indicators to evaluate whether the maximum frequency difference and recovery time meet the requirements. If the verification fails, the configuration scheme is rolled back and adjusted until it passes.For example, in a fault area containing 35 nodes and an initial optimization model allocating 70 MW of fast response capacity, the initial unit node capacity is 2 MW / node. If 10 key topology location nodes are identified in this area and their distribution density indicates a need to increase response capability by 50% at these key nodes, then the unit capacity of these 10 key nodes is adjusted to 3 MW / node, while the remaining nodes remain at 2 MW / node. After weighted calculation, the average unit node capacity in the fault area increases to approximately 2.14 MW / node. If the preset density threshold is 2.0 MW / node, the adjusted value meets the threshold. The system will then issue specific commands and require the corresponding inverters and energy storage systems to respond in a priority manner. If resources are insufficient near some key nodes, the system will automatically trigger neighborhood resource sharing or send a supplementary resource request to the dispatcher.

[0031] Furthermore, based on the multi-factor constrained optimization model, different capacities of fast-frequency response resources are allocated to faulty and non-faulty regions, including: Based on the convergence objective of the optimization function, the complete set of nodes in the fault region is loaded as the spatial domain boundary constraint into the multi-factor constrained optimization model. Based on the correspondence between the identification results of inertial weak points in the complete node set of the fault region and the inertia distribution of the power system, a priority factor for resource allocation in the fault region is generated. Based on the path connectivity characteristics of the network topology, the impact of the priority factor distribution on the transmission attenuation effect in non-faulty areas is verified. Based on the optimization function-driven resource allocation instructions generated by the multi-factor constraint optimization model in response to priority factors and the influence of transmission attenuation effects, the system allocates fast frequency response resource capacity with a dominant optimization level to the fault region, while allocating supplementary capacity resources that conform to the dominant optimization level to the non-fault region.

[0032] As a preferred embodiment of the above, the complete set of nodes in the fault region is first loaded into the multi-factor constrained optimization model as a spatial domain boundary constraint. A preferred method is to explicitly define the minimum guaranteed capacity or minimum guaranteed capacity ratio corresponding to this node set as part of the convergence objective in the optimization problem. This ensures that the optimizer must meet the lower limit requirement of the spatial domain boundary capacity during iterative solution, using this as one of the main constraints to drive convergence. To ensure convergence stability, a phased convergence strategy is preferred. The first stage prioritizes meeting the spatial domain boundary constraints and reducing the maximum node frequency difference for rapid convergence. The second stage refines the region while meeting the boundary constraints. Internal resource allocation is optimized to improve overall performance. A combination of soft constraints and penalty terms is used for boundary constraints to balance solution feasibility and convergence speed. Then, based on the identification of weak points in the fault region node set and the overall power system inertia distribution, a priority factor for fault region resource allocation is generated. The preferred identification method is to compare the equivalent rotational inertia density of each node or node cluster with the regional average inertia density, combined with node importance indicators such as load size, generator connectivity, and topological centrality to define the priority factor. The priority factor represents the node's urgent need for fast frequency response in a qualitative or normalized numerical form; a higher value indicates that the node should... To obtain higher priority resource allocation, historical disturbance experience and real-time PMU frequency drop data are preferred when generating priority factors to ensure the timeliness and representativeness of the factors. Subsequently, based on the path connectivity characteristics of the network topology, the impact of the priority factor distribution on the transmission attenuation effect in non-faulty areas is verified. Preferred verification methods include estimating the impact attenuation of each key node from the faulty area to the non-faulty area based on electrical distance, such as equivalent impedance distance, or coupling degree and path hop count based on power flow sensitivity. If the attenuation is significant on a certain path, it indicates that the high priority allocation in the faulty area has a negligible impact on the non-faulty node; otherwise, the corresponding non-faulty node needs to be adjusted in the optimization model. For fault nodes, a lower limit for response or additional compensation resources are introduced; in practical engineering, it is preferable to set the attenuation criterion as several empirical thresholds and supplement it with online sensitivity analysis to automatically adjust the thresholds to adapt to different disturbance scales; based on the above priority factors and the effect of transmission attenuation, when the resource allocation instruction is generated by the optimization function driven by the multi-factor constraint optimization model, this implementation prefers to adopt a hierarchical priority allocation strategy: first, allocate fast frequency response resource capacity with dominant optimization level to the fault area, that is, reserve and preferentially allocate a certain proportion of high response rate and high power density resources to the fault area in the available resource pool to ensure that the area can locally suppress frequency deviation in the early stage of disturbance;Simultaneously, supplementary capacity resources, conforming to this dominant optimization level, are allocated to non-faulty areas to compensate for and coordinate responses when local resources are insufficient or frequency recovery fails due to cross-regional transmission. To enhance project feasibility, it is preferable to weight the allocation of different nodes within the faulty area according to a priority factor during the allocation process: high-priority inertial weak points receive higher single-node capacity allocation within the same area, while the supplementary capacity of non-faulty areas prioritizes ensuring the overall safety of the system, and after satisfying the priority protection of the faulty area, the remaining resources are distributed according to node importance and connectivity accessibility. To ensure practical effectiveness, this implementation method preferably performs online time-domain verification before generating the final allocation instruction: the obtained resource allocation scheme is input into a fast simulator to evaluate the maximum frequency difference, frequency recovery time, and cross-regional impact. If the verification result fails to reach the convergence target or does not meet the key constraints, the priority factor is automatically adjusted, the protection ratio of the faulty area is increased, or the supplementary capacity of the non-faulty area is adjusted until the simulation passes. In extreme resource-constrained situations, preset emergency remedial measures are triggered first, such as short-term sharing in adjacent areas, emergency release of energy storage devices, or issuance of manual dispatch commands.

[0033] Furthermore, configuring high-frequency response resource capacity with a dominant optimization level to the fault area includes: Spatially map and associate the complete set of nodes in the fault region with the fast frequency response resource capacity of the dominant optimization level. Based on the transitional connection attributes of nodes in the energy transmission path in the network topology, identify the topological connection attribute nodes in the complete node set of the fault area. For nodes with topological connection attributes, topological fluid equilibrium weights are added to the priority factor based on the spatial association algorithm of the multi-factor constraint optimization model. Based on the priority factor with added weight, the fast frequency response resource capacity of the dominant optimization level is redistributed so that the topology connection attribute nodes can obtain a share of configuration capacity with increased proportion. Write the configured capacity share into the execution control parameters for localized suppression of frequency deviation.

[0034] As a preferred embodiment of the above, the set of fault region nodes determined by the multi-factor constrained optimization model is first spatially mapped and associated with the total fast frequency response capacity defined under the dominance optimization level. Preferably, each fault region node is labeled with its allocable capacity limit, initial allocation share, and node importance weight in the power grid topology diagram to facilitate subsequent resource allocation at the node level. Then, based on the transitional connectivity attributes of nodes in the energy transmission path, topological connection attribute nodes are identified. Preferably, topological centrality measures such as betweenness centrality, path traversal count, or power flow-based path flow contribution are combined with actual operational data, such as those observed through the PMU. The frequency of power flow paths and the relay role of converters / substations are used to determine which nodes have connection attributes. Nodes that perform switching or relay functions on multiple main energy transmission paths or connect at least two high-capacity corridors at the regional boundary and undertake significant power flow transfer are preferably identified as nodes with topological connection attributes. After identification, topological fluid equilibrium weights are added to the original priority factors for these nodes according to the spatial association algorithm implemented in the multi-factor constrained optimization model. The preferred addition rule is to integrate the topological fluid equilibrium weights with the node's inertia weakness and node importance weights in a certain proportion and then normalize them to maintain the overall priority factor. To ensure consistent scaling and avoid excessive skewness at single points, it is preferable to set an upper limit for the added weights. For example, the increase in the total weight of a single node is limited to a preset percentage range and can be dynamically scaled according to resource sufficiency. After the weights are added, the fast frequency response capacity under the dominant optimization level is redistributed based on the updated priority factors. A weighted allocation strategy is preferred: first, ensuring that each node receives a capacity no less than its minimum guaranteed capacity and no more than the maximum available capacity of a single device; then, adjusting the allocable margin according to the priority factor ratio, so that nodes with topology connection attributes receive a configuration share with a relative increase in proportion. At the same time, the regional total amount constraint and response rate are strictly observed during the redistribution process. Multiple physical constraints, such as duration, are applied, and real-time robustness checks are performed on potential cross-regional impacts. Finally, the capacity share of each node after reallocation is written into the execution control parameters of localized suppression, preferably including but not limited to the trigger threshold, priority identifier, initial slope and maximum output, response hold time, recovery and yield rules, timestamp and version number, and communication confirmation mechanism for each response unit. These parameters are then distributed to the corresponding energy storage, inverter, and controllable load controllers through existing control and communication mechanisms. Before distribution, it is preferable to verify the effect of the configuration on the maximum frequency difference and recovery time in simulation or small-scale trial operation to identify and handle unforeseen interactive effects.

[0035] Furthermore, such as Figure 4 As shown, the transient bias of the time-domain simulation is corrected based on the correlation between the power angle trajectory and frequency response of the constructed power system nodes, including: Identify the transient power angle swing range of each power system node in the time-domain simulation and obtain the phase amplitude related variables of the power angle trajectory; The phase amplitude-related variables of each power system node are mapped to the oscillatory energy transfer function established for the frequency response, thus generating the angular frequency coupling strength index. The non-monotonic distortion component of the frequency drop depth is corrected based on the power angle frequency coupling strength index. The frequency drop depth after distortion correction is dynamically weighted by the power angle frequency coupling strength index to generate the compensated weighted frequency drop depth. The maximum frequency difference is calculated by using the compensated weighted frequency drop depth and the maximum corrected frequency change rate.

[0036] As a preferred embodiment of the above, firstly, based on the high-precision time-series phase data obtained by the synchronous phasor measurement device, the transient power angle swing interval of each node in the time-domain simulation results is identified, and the phase amplitude-related variables of the power angle trajectory are extracted. The preferred processing flow is as follows: the original phase time series is first subjected to power frequency component suppression and narrowband denoising processing, and then envelope detection is performed within a specified time window after the disturbance is triggered, such as the Hilbert transform method or short-time energy envelope analysis, to extract a set of phase amplitude-related variables such as instantaneous phase amplitude, peak-to-peak value, effective value, and main oscillation mode amplitude, so as to describe the power angle swing characteristics of a single node with multiple indicators; then, the phase amplitude-related variables of each node are mapped to the oscillation energy used to characterize the frequency response. The transfer function generates a power angle and frequency coupling strength index. A preferred mapping method is described as follows: First, perform modal decomposition on the system, such as using the Prandtl method or time-series mode identification based on singular value decomposition, to obtain the main oscillation modes and their modal participation factors. Combine the phase amplitude at the node and the system's local equivalent moment of inertia information to calculate the node's kinetic energy contribution in each mode. Then, accumulate the contributions of each mode according to their energy meaning and normalize them to obtain a coupling strength index in the range of 0 to 1. A larger value indicates that the node's power angle oscillation is more likely to transfer energy to the entire network's frequency response. Based on this coupling strength index, correct the non-monotonic distortion components of the simulated frequency drop depth. Specifically, the preferred method is to first identify... Non-monotonic features in the frequency curve caused by numerical values, models, or boundary conditions, such as short-term multiple local extrema or jitter that does not conform to the energy conservation expectation, are addressed by determining the energy distribution that the ideal drop pattern of the node should present based on the coupling strength. Local extrema identified as distortions are replaced by envelope reconstruction weighted by modal energy distribution or smooth reconstruction based on the minimum norm, so that the corrected frequency drop curve is consistent with the power angle trajectory in terms of overall energy and modal participation. Subsequently, a dynamic time constant weighting is applied to the distortion-corrected frequency drop depth according to the coupling strength to generate a compensated weighted drop depth. The preferred weighting rule is to map the coupling strength to a set of time constant factors, with higher coupling strengths corresponding to shorter time constants. The time constant is used to reflect the instantaneous drop caused by rapid energy transfer, while the time constant is longer for coupling strength to reflect the slow energy accumulation effect. In practice, a smooth monotonic mapping function is used and upper and lower limits are applied to extreme values ​​to avoid overcorrection. At the same time, the local inertia and damping characteristics of the nodes are considered in the weighting process to correct the actual physical meaning of the time constant. Finally, the frequency drop depth after compensation weighting and the maximum frequency change rate obtained by the same correction process are used together to calculate the maximum frequency difference of the node. A conservative value selection strategy is preferred, that is, when there is uncertainty, the combination value that can reflect the most unfavorable working condition is selected, and if necessary, the interval estimate is obtained by multiple parameter perturbation simulations for use in the optimization model.In engineering applications, it is recommended to run this correction process and a set of small parameter perturbations in parallel for each fault event to obtain robust post-compensation frequency indices, which can then be used as input for subsequent partition allocation optimization.

[0037] Furthermore, adjusting the resource capacity of each node in the faulty region includes: Identify the topological location attributes of each node in the complete node set of the fault area, and output the location configuration coordinates of key topological location nodes; Calculate the connectivity distribution index of key topological location nodes based on location configuration coordinates; Gradient adjustment calculations are performed on the allocation of resource capacity per unit node based on connectivity distribution indicators; Verify whether the unit node resource capacity after gradient adjustment operation enhances the node distribution density advantage in the fault region; The aggregation effect of the enhanced unit node resource capacity on the complete set of nodes in the fault region is statistically analyzed and used as the input resource allocation scheme for the adjusted unit node resource capacity.

[0038] As a preferred embodiment of the above, after identifying the set of nodes in the fault area, the location coordinates of key topological nodes are output based on network topology information and available geographical location information or topological location mapping generated by a graph layout algorithm. A preferred approach is to directly use latitude and longitude or site coordinates as location coordinates when geographical data is available; when only topology information is available, force-directed or spectral layout methods are used to generate planar coordinates to reflect the spatial relationships of nodes in the energy transmission network, and the line capacity, line impedance, and adjacency of each node are recorded for subsequent calculations. Then, key topological nodes are calculated based on these location coordinates and topology attributes. The connectivity distribution index of a node is preferably defined as a composite evaluation result of multiple connectivity measures. This composite evaluation includes node degree (number of direct neighbors), weighted degree (considering line rated capacity and power flow weight), betweenness centrality (measuring the frequency of a node acting as a path relay), proximity (average topological distance to other nodes), and local neighborhood density (number of nodes and total capacity within a given electrical distance or topological radius). Each measure is normalized and weighted to obtain the connectivity distribution index value for each key node. When performing gradient adjustment calculations on the allocation of resource capacity per unit node based on the obtained connectivity distribution index, this implementation method... The preferred approach employs an iterative gradient allocation strategy: using the initial unit node capacity as a baseline, the direction and magnitude of capacity adjustment are determined by the connectivity distribution index. Specifically, more available fast frequency response capacity is proportionally allocated to nodes with high connectivity that are at key energy transmission points or boundary connections in the topology. Simultaneously, upper and lower limits are set for the adjustment increment to avoid excessive concentration at a single point. A neighborhood smoothing mechanism is introduced in each iteration, ensuring that the capacity adjustment of a node depends not only on its own connectivity index but also on the indices of its neighboring nodes, thus forming a gradient-like, progressively strengthening resource distribution. Furthermore, during the iteration process, the total capacity constraints of the entire region and the approximate capacity of individual nodes are checked in real time. The constraints and response characteristics are met; after adjustment, verification is performed to determine whether the unit node resource capacity after gradient adjustment has enhanced the node distribution density advantage in the fault area. Preferably, the enhancement effect is quantified by calculating the average unit node capacity, capacity density change per unit area or per unit electrical radius in the same area before and after adjustment, and the degree of imbalance, such as the coefficient of variation or clustering index, and these measures are compared with the preset density advantage threshold. If the verification fails, the gradient step size is increased or decreased, the neighborhood smoothing intensity is adjusted, or the key node weights are reselected according to the preferred strategy, and the iteration is repeated until the threshold is met or the acceptable resource utilization boundary is reached.Finally, the aggregation effect of the enhanced unit node resource capacity on the faulty node set is statistically analyzed and a summary report is generated. This report serves as the final input to the resource allocation scheme for the adjusted unit node resource capacity. The statistics include the final allocated capacity distribution for each node, the percentage of incremental capacity obtained by key nodes, the regional average unit node capacity, and its confidence interval. These results are then incorporated into subsequent configuration commands and simulation verification processes for final verification before execution.

[0039] Example 2; Based on the same inventive concept as the frequency modulation resource partitioning allocation method considering frequency distribution in the foregoing embodiments, the present invention also provides a frequency modulation resource partitioning allocation system considering frequency distribution, the system comprising: The information acquisition module acquires the frequency dynamic indicators of each power system node in the power system and calculates the maximum frequency difference of the power system nodes; The region division module dynamically divides the power system into fault regions and non-fault regions based on the location of the disturbance source. The fault region includes the disturbance source node and a set of nodes within a preset electrical distance threshold. The function optimization module establishes an optimization function with the objective of minimizing the maximum value among the maximum frequency differences of power system nodes; The resource allocation module, based on an optimization function, allocates fast frequency response resources of different capacities to faulty and non-faulty areas through a multi-factor constrained optimization model according to the maximum frequency difference of power system nodes, power system inertia distribution, network topology, and electrical distance parameters. The local suppression module configures fast frequency response resources with a density higher than a preset threshold in the fault area to achieve localized suppression of frequency deviation.

[0040] The adjustment system described above in this invention can effectively implement a frequency modulation resource partitioning allocation method that takes into account frequency distribution, and the technical effects it can achieve are as described in the above embodiments, and will not be repeated here.

[0041] Furthermore, the information acquisition module includes: The time-domain simulation unit, based on real-time measurement data from synchronous phasor measurements, obtains the maximum frequency change rate and frequency drop depth of power system nodes as dynamic frequency indicators during the time-domain simulation process. The deviation correction unit corrects the transient deviation of the time-domain simulation based on the correlation between the power angle trajectory and frequency response of the constructed power system nodes. The frequency difference calculation unit calculates the maximum frequency difference of power system nodes based on the corrected maximum frequency change rate and frequency drop depth.

[0042] Similarly, the above-mentioned optimization schemes for the system can also achieve the optimization effects corresponding to the methods in Embodiment 1, which will not be repeated here.

[0043] Although this application has been described in conjunction with specific features and embodiments, it is obvious that various modifications and combinations can be made thereto without departing from the spirit and scope of this application. Accordingly, this specification and drawings are merely exemplary illustrations of the application as defined herein, and are to be considered as covering any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Thus, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A frequency resource partitioning allocation method considering frequency distribution, characterized in that, The method comprises: acquiring frequency dynamic indicators of each power system node in a power system and calculating maximum frequency deviation of the power system node; dividing the power system dynamics into a fault area and a non-fault area according to the position of the disturbance source, wherein the fault area comprises the disturbance source node and a node set within a preset electrical distance threshold range; establishing an optimization function aiming to minimize the maximum value in the maximum frequency deviation of the power system node; based on the optimization function, allocating different capacity of fast frequency response resources to the fault area and the non-fault area according to the maximum frequency deviation of the power system node, power system inertia distribution, network topology and electrical distance parameters through a multi-factor constraint optimization model; configuring fast frequency response resources with a density higher than a preset threshold in the fault area to achieve localized suppression of frequency deviation.

2. The method for frequency distribution taking into account the allocation of frequency resources partition according to claim 1, characterized in that, The method comprises: acquiring frequency dynamic indicators of each power system node in a power system and calculating maximum frequency deviation of the power system node; based on real-time measurement data of synchronous phasor measurement, acquiring the maximum frequency rate and frequency drop depth of the power system node in a time domain simulation process as the frequency dynamic indicators; correcting transient deviation of the time domain simulation according to the correlation between the power angle trajectory and frequency response of the power system node; 3. The method of claim 1, wherein, calculating the maximum frequency deviation of the power system node based on the corrected maximum frequency rate and frequency drop depth. The method comprises: calculating distance relationship metrics of all the power system nodes and the disturbance source node based on the electrical distance parameters; according to the comparison result of the value of the distance relationship metrics and the preset electrical distance threshold, screening out nodes with an electrical coupling strength reaching a preset requirement to form an initial candidate set; based on the adjacency matrix of the network topology, verifying the reachability of each node in the initial candidate set to the disturbance source node according to the topological path, and excluding nodes not satisfying the connectivity constraint; 4. The method for frequency distribution taking into account the allocation of frequency resources partition according to claim 3, characterized in that, integrating the verified nodes and the disturbance source node to form a complete node set of the fault area. The method comprises: acquiring capacity allocation indicators of the fault area and the non-fault area output by the multi-factor constraint optimization model; calculating unit node resource capacity of the fault area and the non-fault area; based on the distribution density of key topological position nodes in the complete node set of the fault area, adjusting the unit node resource capacity of the fault area; 5. The method for frequency distribution taking into account the allocation of frequency resources partition according to claim 3, characterized in that, according to the comparison result of the adjusted unit node resource capacity of the fault area and the unit node resource capacity of the non-fault area, confirming a resource allocation scheme satisfying the density higher than the preset threshold, and outputting a final configuration instruction based on the localized suppression of the fast frequency response resources. The method comprises: loading the complete node set of the fault area as a spatial domain boundary constraint condition to the multi-factor constraint optimization model based on a convergence target of the optimization function; generating a priority factor of resource allocation of the fault area based on a corresponding relationship between an inertial weak point identification result in the complete node set of the fault area and an inertia distribution of the power system; verifying an influence transmission attenuation effect of the priority factor on the non-fault area according to a path connectivity feature of the network topology; based on the optimization function, driving resource allocation instructions generated according to the multi-factor constraint optimization model in response to the priority factor and the influence transmission attenuation effect, configuring the fault area with fast frequency response resource capacity of a dominant optimization level, and configuring the non-fault area with supplementary capacity resources subject to the dominant optimization level.

6. The method for frequency distribution taking into account the allocation of frequency resources partition according to claim 5, characterized in that, configuring the fault area with fast frequency response resource capacity of a dominant optimization level, comprising: spatially mapping and associating the complete node set of the fault area with the fast frequency response resource capacity of the dominant optimization level; based on the transition connection attribute of nodes in the energy transmission path in the network topology, identifying a topological connection attribute node in the complete node set of the fault area; for the topological connection attribute node, appending a topological fluid balance weight to the priority factor according to a spatial association algorithm of the multi-factor constraint optimization model; redistributing the fast frequency response resource capacity of the dominant optimization level according to the priority factor with the appended weight, so that the topological connection attribute node obtains a configuration capacity share with an increased proportion; writing the configuration capacity share into the execution control parameter of the localized suppression of the frequency deviation.

7. The method for frequency distribution taking into account the allocation of frequency resources partition according to claim 2, characterized in that, correcting the transient deviation of the time domain simulation according to the correlation between the power angle trajectory and the frequency response of the constructed power system node, comprising: identifying the transient power angle swing interval of each power system node in the time domain simulation and obtaining the phase amplitude related variable of the power angle trajectory; mapping the phase amplitude related variable of each power system node to an oscillation energy transfer function established for frequency response, generating a power angle frequency coupling strength index; correcting the non-monotonicity distortion component of the frequency drop depth based on the power angle frequency coupling strength index; performing dynamic time constant weighting on the frequency drop depth after distortion correction according to the power angle frequency coupling strength index, generating a compensation weighted frequency drop depth; using the compensation weighted frequency drop depth and the corrected maximum frequency change rate to calculate the maximum frequency difference.

8. The method for frequency distribution taking into account the allocation of frequency resources partition according to claim 4, characterized in that, adjusting the unit node resource capacity of the fault area, comprising: identifying the topological position attribute of each node in the complete node set of the fault area, outputting the position configuration coordinates of the key topological position node; calculating the connectivity distribution index of the key topological position node based on the position configuration coordinates; performing gradient adjustment operation on the unit node resource capacity allocation according to the connectivity distribution index; verify whether the unit node resource capacity after the gradient adjustment operation enhances the node distribution density advantage of the fault area; statistically aggregate the effect of the enhanced unit node resource capacity on the complete node set of the fault area, and input the adjusted unit node resource capacity into the resource allocation scheme.

9. A frequency distribution aware frequency resource partitioning allocation system, characterized by, The system comprises: an information acquisition module that acquires frequency dynamic indicators of each power system node in a power system and calculates the maximum frequency difference of the power system node; a region division module that divides the power system into a fault area and a non-fault area according to the location of the disturbance source, wherein the fault area includes the disturbance source node and a node set within a preset electrical distance threshold; a function optimization module that establishes an optimization function with the minimum value of the maximum frequency difference of the power system node as the target; a resource allocation module that allocates different capacity of fast frequency response resources to the fault area and the non-fault area based on the optimization function, the maximum frequency difference of the power system node, the inertia distribution of the power system, the network topology, and the electrical distance parameters, and through a multi-factor constraint optimization model; a local suppression module that configures fast frequency response resources with a density higher than a preset threshold in the fault area to achieve local suppression of frequency deviation.

10. The frequency distribution taking frequency resource partitioning allocation system of claim 9, wherein, The information acquisition module comprises: a time domain simulation unit that acquires the maximum frequency rate and the frequency drop depth of the power system node as the frequency dynamic indicators in the time domain simulation process based on the real-time measurement data of the synchronized phasor measurement; a deviation correction unit that corrects the transient deviation of the time domain simulation according to the correlation between the power angle trajectory and the frequency response of the power system node; a frequency difference calculation unit that calculates the maximum frequency difference of the power system node based on the corrected maximum frequency rate and the frequency drop depth.

Images