Islanding method and system based on topology reconfiguration and key load

Abstract

The application discloses a power grid island division method and system based on topology reconstruction and key load, relates to the technical field of power system safety regulation, and comprises the following steps: spatially coupling typhoon wind field data and geographical space information of each element in a power grid model to obtain the wind field action intensity borne by each element at each moment; determining the failure risk of each element according to the wind field action intensity based on an element damage evaluation model; determining the failure elements that exit operation along the time sequence of typhoon advancement, reconstructing the topology connection relationship of the power grid model, performing power flow calculation, and obtaining the failure evolution deduction result; constructing an island division objective function with the maximum load recovery degree as the target, solving the island division objective function based on the reconstructed topology connection relationship, and outputting the power grid island division strategy. The application has the beneficial effects that the power grid model is adapted to the actual operable element state, and the island division strategy is generated in advance, and the timeliness of the power grid island division is considered.

Description

Technical Field

[0001] This application relates to the field of power system security control technology, and in particular to a method and system for grid islanding based on topology reconfiguration and critical loads. Background Technology

[0002] Coastal areas are prone to typhoon disasters. The extreme wind speeds, heavy rainfall, and storm surges brought by typhoons pose a serious threat to the safe operation of power infrastructure. In recent years, with the continuous growth of electricity load in coastal economic zones, the power grid has expanded in scale and become increasingly complex in topology. The economic losses and social impacts caused by large-scale power outages due to typhoons have become increasingly significant. During typhoons, transmission line collapses, substation equipment damage, and other faults often occur simultaneously, leading to drastic changes in the power grid topology and the loss of power to a large number of loads. Accurately assessing the damage to the power grid and developing effective power restoration strategies under typhoon conditions to ensure continuous power supply to critical loads such as hospitals, communication base stations, and emergency command centers is a crucial technical problem that urgently needs to be solved in the field of power system disaster prevention and mitigation.

[0003] Currently, the technical means for power grid safety analysis under typhoon disasters mainly rely on early warning information issued by meteorological departments. Meteorological early warning data typically includes macro-meteorological elements such as typhoon path prediction, central pressure, maximum wind speed, and affected area, providing power grid dispatchers with an overall assessment of the situation before the disaster strikes. However, this type of macro-meteorological information is difficult to directly map onto specific physical components of the power grid. Power grid components such as transmission lines, towers, and substations are distributed across a vast geographical area. Components of different locations, types, and design standards exhibit significant differences in their ability to withstand wind disasters. Relying solely on regional wind speed levels or typhoon path information is insufficient to accurately determine which specific components will face failure risks and when, resulting in a lack of refined support for dispatch decisions based on specific lines and equipment.

[0004] In power grid fault analysis, existing methods primarily rely on offline analysis based on pre-defined static fault sets. Typical Nk fault analysis assumes that faulty components are a pre-given or one-time determined set, and evaluates the system's response and recovery capabilities accordingly. However, typhoon disasters exhibit significant spatial propagation and temporal continuity. Typhoons gradually affect power grid facilities in different regions along their path, and the failure of power grid components does not occur simultaneously but exhibits a clear temporal sequence as the typhoon progresses. Static fault set analysis cannot reflect this dynamic process of gradual advancement along geographical space leading to the sequential failure of power grid components. Therefore, it is difficult to accurately characterize the propagation path of faults and the evolution of system status under disaster conditions, resulting in a failure to balance adaptability and timeliness in the division of power grid islands. Summary of the Invention

[0005] This application addresses the problem of the inability to simultaneously achieve adaptability and timeliness in existing power grid islanding methods. It provides a power grid islanding method and system based on topology reconfiguration and critical loads. By assessing the failure risk of components under wind field intensity, high-probability failure components are removed from the power grid model, and the power grid model is reconstructed to ensure that the power grid model always adapts to the actual operational component states. Simultaneously, the fault is extrapolated along the typhoon progression timeline. When the fault evolution extrapolation results meet the power grid islanding triggering conditions, the power grid islanding strategy is immediately calculated based on the islanding objective function. The generated islanding strategy prioritizes the restoration of critical loads while actively avoiding power supply paths with high failure risks, preventing secondary disconnection due to power supply path failures after islanding is formed. Furthermore, the islanding strategy is pre-generated to ensure the timeliness of power grid islanding.

[0006] To achieve the aforementioned technical objectives, this application provides a technical solution: a power grid islanding method based on topology reconfiguration and critical loads, comprising the following steps: constructing a power grid model containing topological connections, geospatial information, and electrical operating parameters based on power grid structure data; acquiring typhoon wind field data and spatially coupling the typhoon wind field data with the geospatial information of each component in the power grid model to obtain the wind field intensity experienced by each component at each time; determining the failure risk of each component based on the wind field intensity using a component damage assessment model; determining the faulty components that will be taken out of operation based on the failure risk along the typhoon progression sequence, and reconstructing the topological connections of the power grid model; performing power flow calculations on the reconstructed topological connections to obtain fault evolution simulation results; calculating the load recovery degree based on the critical load recovery amount and the power supply path failure risk penalty; constructing an islanding objective function with the goal of maximizing the load recovery degree; and when the fault evolution simulation results meet the power grid islanding triggering conditions, solving the islanding objective function based on the reconstructed topological connections to output the power grid islanding strategy.

[0007] Furthermore, the construction of a power grid model based on power grid structure data, which includes topological connectivity, geospatial information, and electrical operating parameters, includes: constructing topological connectivity with power grid nodes as vertices of the graph model and power grid components as branches of the graph model; assigning electrical operating parameters to each branch and geospatial information to each vertex and branch; and geometricating components with spatial extension attributes into spatial line segments.

[0008] Furthermore, the spatial coupling of typhoon wind field data with the geospatial information of each component in the power grid model includes: performing spatial intersection calculations on the typhoon wind field data with the geospatial information of the spatial line segments and each vertex in the power grid model at each moment, and extracting the wind speed at the location of each component as the wind field intensity; for a spatial line segment that crosses multiple spatial grid units of the typhoon wind field data, taking the maximum value of the wind speed in each spatial grid unit that the spatial line segment passes through or the average value weighted by the component length as the wind field intensity of the component corresponding to the spatial line segment.

[0009] Furthermore, the component damage assessment model employs a logistic function, with the critical wind resistance of the component serving as the inflection point parameter of the logistic function.

[0010] Furthermore, the step of determining the out-of-operation faulty components based on the failure risk along the typhoon progression sequence and reconstructing the topology connection relationship of the power grid model includes: at each moment of the typhoon progression sequence, when the component failure risk meets the preset fault judgment condition, determining the component as an out-of-operation faulty component; in response to the switch state change caused by the out-of-operation of the faulty component, updating the connectivity relationship between components, deleting the out-of-operation branches, and reconstructing the topology connection relationship of the power grid model.

[0011] Furthermore, the step of performing power flow calculations on the reconstructed topology connections to obtain fault evolution prediction results includes: performing connected component search on the reconstructed topology connections to identify electrical islands and the power supply and load distribution of the electrical islands; performing power flow calculations on each electrical island to obtain voltage-over-limit buses, power flow-over-limit branches, and power-loss loads as fault evolution prediction results.

[0012] Furthermore, the calculation of the load recovery degree based on the critical load recovery amount and the power supply path failure risk penalty includes: obtaining the load recovery amount weight based on the critical priority of the load, calculating the critical load recovery amount with the load recovery amount weight and each load recovery amount; calculating the power supply path failure risk penalty according to the failure risk of the branch; and calculating the load recovery degree with the difference between the critical load recovery amount and the power supply path failure risk penalty.

[0013] Furthermore, the calculation of the load recovery degree based on the critical load recovery amount and the power supply path failure risk penalty also includes: calculating the action loss based on the switching operation cost and switching action status of the branch; and calculating the load recovery degree by subtracting the action loss and the power supply path failure risk penalty from the critical load recovery amount.

[0014] Furthermore, the islanding strategy for the power grid is output by solving the islanding objective function based on the reconstructed topology connection relationship includes: solving the islanding objective function based on the reconstructed topology connection relationship to output candidate islanding schemes; performing AC power flow calculation and frequency response verification on the candidate islanding schemes; if the verification fails, generating cut constraints based on the limit information, incorporating the cut constraints into the constraints of the islanding objective function, and re-executing the islanding objective function solution until the verification passes or the set maximum number of iterations is reached.

[0015] Another technical solution provided in this application is a power grid islanding system based on topology reconfiguration and critical loads, used to implement the method described above, including: a model building unit, used to build a power grid model containing topological connections, geospatial information and electrical operating parameters based on power grid structure data; a wind field assessment unit, used to spatially couple typhoon wind field data with the geospatial information of each component in the power grid model to obtain the wind field intensity experienced by each component at each time; a failure assessment unit, used to determine the failure risk of each component based on the component damage assessment model according to the wind field intensity; a topology reconfiguration unit, used to determine the faulty components that will be taken out of operation according to the failure risk along the typhoon progression sequence, and reconfigure the topology connections of the power grid model; a fault inference unit, used to perform power flow calculation on the reconfigured topology connections to obtain fault evolution inference results; and an islanding unit, which pre-stores an islanding objective function, used to solve the islanding objective function based on the reconfigured topology connections and output the power grid islanding strategy when the fault evolution inference results meet the power grid islanding triggering conditions.

[0016] The beneficial effects of this application are as follows: By pre-constructing a power grid model and following the progress of a typhoon, the intensity of wind field action on different components during the typhoon's progression is obtained. The failure risk of components under the wind field intensity is assessed, and faulty components with a high probability of failure are removed from the power grid model. The power grid model is then reconstructed, ensuring that it always adapts to the actual operational component states. Faults are then simulated. When the fault evolution simulation results meet the triggering conditions for power grid islanding, the power grid islanding strategy is immediately calculated based on the islanding objective function. The generated islanding strategy prioritizes the restoration of critical loads while actively avoiding power supply paths with high failure risks, preventing secondary disconnection due to power supply path failures after islanding is formed. Furthermore, since the fault simulation unfolds along the typhoon's progression, the damaged topology can be predicted in the future before the typhoon's actual impact arrives, and an islanding strategy can be generated in advance, ensuring the timeliness of islanding. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the grid islanding method based on topology reconfiguration and critical loads in this application.

[0018] Figure 2 This is a schematic diagram illustrating the solution process of the grid islanding strategy in the grid islanding method based on topology reconfiguration and critical loads in this application. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description of this application is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely one preferred embodiment of this application and are only used to explain this application. They do not limit the scope of protection of this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0020] like Figure 1 As shown, the grid islanding method based on topology reconfiguration and critical loads includes the following steps: A power grid model is constructed based on power grid structure data, including topological connections, geospatial information, and electrical operating parameters. Acquire typhoon wind field data, and spatially couple the typhoon wind field data with the geospatial information of each component in the power grid model to obtain the wind field intensity experienced by each component at each time. The failure risk of each component is determined based on the intensity of the wind field effect, according to the component damage assessment model. Based on the failure risk advancing along the typhoon timeline, the faulty components that will be taken out of service are determined, and the topological connections of the power grid model are reconstructed. Perform power flow calculations on the reconstructed topology connections to obtain fault evolution prediction results; The degree of load recovery is calculated based on the critical load recovery amount and the penalty for power supply path failure risk. An islanding objective function is constructed with the goal of maximizing load recovery. When the fault evolution simulation results meet the grid islanding triggering conditions, the grid islanding strategy is output by solving the islanding objective function based on the reconstructed topology connection relationship.

[0021] In this embodiment, a pre-constructed power grid model is used to track the typhoon's progress and obtain the wind field intensity experienced by different components during the typhoon's advance. The failure risk of components under the wind field intensity is assessed, and high-probability failure components are removed from the power grid model. The power grid model is then reconstructed to ensure that it always adapts to the actual operational component states. Fault prediction is then performed, and when the fault evolution prediction results meet the power grid islanding trigger conditions, the power grid islanding strategy is immediately calculated based on the islanding objective function. This generated islanding strategy prioritizes the restoration of critical loads while actively avoiding power supply paths with high failure risks, preventing secondary disconnection due to power supply path failures after islanding is formed. Furthermore, since the fault prediction unfolds along the typhoon's advance sequence, the damaged topology can be predicted in the future before the typhoon's actual impact arrives, and an islanding strategy can be generated in advance, ensuring the timeliness of islanding.

[0022] Specifically, constructing a power grid model based on power grid structure data, including topological connectivity, geospatial information, and electrical operating parameters, includes: The topological connection relationship is constructed by using power grid nodes as vertices of the graph model and power grid components as branches of the graph model. Electrical operating parameters are assigned to each branch, and geospatial information is assigned to each vertex and branch. Elements with spatial extension properties are geometricated into spatial line segments.

[0023] Using the busbars or topology nodes in the power grid as the graph model vertex set Branch circuits are composed of components such as transmission lines and transformers. Establish the topological connections of the power grid model. For each branch... Additional electrical operating parameters, including line resistance. Reactance Voltage to ground This includes the impedance of each transformer winding and its active power, current, or apparent power operating limits. Geographical spatial information is assigned to each vertex and branch. For overhead transmission lines, based on the latitude and longitude coordinates of their two end towers or substations, the overhead transmission line is geometrically transformed into a spatial line segment with a start and end point and a direction. For power plants such as substations, their latitude and longitude coordinates are used to represent spatial points. .

[0024] The spatial line segment can be further subdivided into several segments according to the tower location, so that the wind field intensity borne by each span can be calculated segment by segment in subsequent steps, thereby improving the spatial accuracy of failure risk calculation.

[0025] Spatial coupling of typhoon wind field data with the geospatial information of various components in the power grid model includes: At each moment, the typhoon wind field data is spatially intersected with the geospatial information of each vertex in the spatial line segment and power grid model, and the wind speed at the location of each component is extracted as the wind field intensity. For a spatial line segment that spans multiple spatial grid cells of typhoon wind field data, the maximum wind speed or the average wind speed weighted by element length in each spatial grid cell through which the spatial line segment passes is taken as the wind field intensity of the element corresponding to the spatial line segment.

[0026] Specifically, typhoon wind field data is typically a time-updated wind speed grid field. Typhoon wind field data is obtained from the output of a mesoscale numerical weather prediction model and after downscaling.

[0027] At each moment, the wind speed grid field is spatially intersected with the spatial line segments and the geospatial information of the power plant to extract the wind speed at the location of each component, thus obtaining the component's... At any moment The representative wind speed that can be withstood .

[0028] In this embodiment, the calculation method for wind field intensity is selected based on the actual partitioning requirements. If the actual partitioning requirement prioritizes safety, then for a spatial line segment that crosses multiple spatial grid cells of the typhoon wind field data, the maximum wind speed in each spatial grid cell traversed by the spatial line segment is taken as the wind field intensity of the element corresponding to that spatial line segment. If the actual partitioning requirement prioritizes overall control, then for a spatial line segment that crosses multiple spatial grid cells of the typhoon wind field data, the weighted average value based on the element length is taken as the wind field intensity of the element corresponding to that spatial line segment.

[0029] By spatially coupling, meteorological grid data and power grid models are precisely aligned in geospatial space, enabling subsequent wind field intensity extraction and failure assessment to be quantitatively analyzed based on the actual wind load borne by each component, thereby improving the accuracy of failure assessment.

[0030] In this embodiment, the component damage assessment model uses the logistic function, with the critical wind resistance of the component as the inflection point parameter of the logistic function.

[0031] Specifically, it will represent wind speed Substitute the pre-established component damage assessment model to calculate the component. At any moment failure risk : ; in, For components Critical wind resistance strength, Characterization element The coefficient of the steepness of the failure probability as the wind speed increases.

[0032] It should be noted that the critical wind resistance strength of a component, i.e., the critical wind speed, varies depending on the component's design wind resistance level.

[0033] In other embodiments, a component damage assessment model may also be constructed using a piecewise linear function, a Weibull distribution function, or other functional forms that can describe the monotonically increasing relationship between wind speed and failure risk.

[0034] Based on the failure risk advancing along the typhoon's timeline, the faulty components that will be taken out of service are identified, and the topology connections of the power grid model are reconstructed, including: At each moment of the typhoon propulsion sequence, when the risk of component failure meets the preset fault judgment conditions, the component is determined to be a faulty component that is taken out of operation. In response to the switch state changes caused by the exit of a faulty component, the connectivity between components is updated, the exited branches are deleted, and the topology of the power grid model is reconstructed.

[0035] As the typhoon progresses, the power grid topology undergoes a dynamic evolution. At time t1, all branches operate normally; at time t2, a certain line is identified as a faulty component due to its failure risk meeting preset fault judgment conditions. The system automatically disconnects the switches on both sides of the faulty line and removes the branch from the power grid model, resulting in a change in topology connectivity; at time t3, more components fail successively, forming new power outage areas. This time-series dynamic reconfiguration mechanism reflects the propagation path and evolution of faults under typhoon disasters.

[0036] In this embodiment, the preset fault determination conditions include: A failure risk threshold is pre-set. At each moment of the typhoon propulsion sequence, the failure risk of each component is compared with this threshold. When the failure risk of a component exceeds the threshold, the component is determined to be out of service; or, Monte Carlo random sampling of component states is performed using the failure risk of each component as the probability. Components selected as failure states are determined to be out of operation, and the probability distribution of fault evolution is obtained through multiple sampling statistics.

[0037] Perform power flow calculations on the reconstructed topology connections to obtain fault evolution projection results, including: The reconstructed topology is searched for connected components to identify electrical islands and their power and load distribution. For each electrical island, perform power flow calculations to obtain the voltage over-limit bus, power flow over-limit branch, and power failure load, which are used as the results of fault evolution prediction.

[0038] After each topology reconfiguration, a graph search algorithm (such as depth-first search or breadth-first search) is first used to identify the current connected components, i.e., electrical islands. For each electrical island, it is checked whether it contains at least one power source node (such as a DG node). If a power source exists, power flow calculation is performed to verify whether the bus voltage is within the allowable range and whether the branch power flow exceeds the thermal stability limit; if no power source exists or the power flow calculation does not converge, the loads within the electrical island are determined to be power-off loads. These limit-crossing and power-off information constitute the fault evolution inference results, which serve as the basis for determining whether islanding needs to be initiated and as the boundary conditions for generating subsequent islanding strategies.

[0039] The triggering conditions for grid islanding include at least the following: the representative wind speed around the substation exceeds the set threshold, or the main grid is disconnected, or the load loss exceeds the set proportion.

[0040] In this embodiment, calculating the load recovery level based on the critical load recovery amount and the power supply path failure risk penalty includes: The load recovery amount weight is obtained based on the key priority of the load, and the key load recovery amount is calculated using the load recovery amount weight and each load recovery amount. Calculate the power supply path failure risk penalty based on the branch failure risk; The degree of load recovery is calculated based on the difference between the critical load recovery amount and the penalty for power supply path failure risk.

[0041] By deducting the power supply path failure risk penalty from the critical load restoration amount, the load restoration level simultaneously reflects both the scale of critical load restoration and the future failure risk of the power supply path. Failure risk is utilized a second time here, its purpose extending from determining whether a component has failed to assessing the future failure probability of candidate power supply paths. This enables the generated islanding strategy to proactively avoid high-risk power supply paths while prioritizing the restoration of critical loads, thus possessing the ability to withstand secondary failures.

[0042] In other embodiments, calculating the load recovery level based on the critical load recovery amount and the power supply path failure risk penalty further includes: Calculate the action loss based on the switching operation cost of the branch and the switching action state; The degree of load recovery is calculated by subtracting the action loss and the penalty for power supply path failure from the critical load recovery amount.

[0043] At this point, based on the reconstructed topology, the following decision variables are defined: load input. , indicating load Whether it is included in the electrified island; branch circuit is energized. , indicating a branch Whether it can handle isolated power supply; and the output of each power source. The objective function for island partitioning is: ; in, Represents the set of branches. For load set, For load The load restoration weight is determined based on the critical priority of the load. For load The active power demand; branch road The corresponding switching operation cost, A 0-1 variable representing the switch's operating state, i.e., whether the switch is activated; branch road failure risk ; , This is a weighting factor.

[0044] Understandably, the critical priority of loads can be determined based on the actual load capacity or set by experts. The trade-off coefficient can be dynamically adjusted according to the typhoon intensity level; the higher the typhoon intensity, the larger the trade-off coefficient corresponding to the penalty for power path failure.

[0045] The power supply path failure risk penalty makes the optimization results tend to exclude branches with high failure probability from the island power supply path, thereby improving the ability of the generated island to resist secondary failures as the typhoon continues to advance.

[0046] In this embodiment, a mixed-integer linear programming model is constructed based on the island partitioning objective function and constraints. The process of solving the island partitioning objective function is the process of solving the mixed-integer linear programming model.

[0047] Specifically, the constraints should include at least the following: Power balance constraints for each island Its internal power output With the input load The power deviation between them shall not exceed the allowable value. : ; Among them, power output Less than or equal to the maximum power output, power output Greater than or equal to the minimum power output; Branch power flow and capacity constraints: The active power flow of branches is represented by DC power flow linearization, and it is constrained to not exceed the branch capacity limit. : ; Frequency feasibility constraints are used to estimate frequency offsets based on the power imbalance of each island. And impose restrictions: ; in, For the isolated island The system power-frequency response coefficient; For the isolated island The power imbalance; This refers to the allowable limit for frequency offset; Connectivity constraints ensure that each charged island is topologically connected and contains at least one power source. Connectivity constraints are modeled using single-item flow or multi-item flow methods. Critical load priority constraints require that, within the limits of power and operational constraints, critical loads with a priority higher than a set level must be included in live islanding; that is, for such loads... .

[0048] like Figure 2 As shown, the islanding strategy for the power grid is as follows: Based on the reconstructed topology connectivity, the objective function for islanding is solved. Based on the reconstructed topological connectivity, the objective function for island partitioning is solved, and candidate island partitioning schemes are output. Perform AC power flow calculations and frequency response verifications on the candidate islanding schemes; If the verification fails, cut constraints are generated based on the out-of-limit information. These cut constraints are then incorporated into the constraints of the island partitioning objective function. The island partitioning objective function is then solved again until the verification passes or the maximum number of iterations is reached.

[0049] This algorithm employs an iterative approach with upper-level solution and lower-level verification. The upper layer rapidly outputs candidate islanding partitioning schemes based on a linearized model, while the lower layer accurately verifies the feasibility of these schemes through AC power flow calculations and frequency response checks. When the lower-level verification detects voltage, frequency, or branch power flow exceeding limits, it generates cut constraints based on this information and feeds them back to the upper-level model. This allows the upper-level model to eliminate infeasible schemes that lead to these limits in subsequent solutions. The algorithm iterates repeatedly until a candidate scheme passes verification or reaches the set maximum number of iterations. This approach balances solution efficiency and computational accuracy, avoiding the frequency and voltage limit exceeding issues that can occur in practical applications with purely linearized models. Furthermore, the maximum number of iterations ensures the convergence of the algorithm, making the final islanding partitioning strategy practically executable.

[0050] Specifically, in the upper-level solution process, a mixed-integer linear programming solver is invoked to solve the mixed-integer linear programming model, obtaining the candidate boundaries for island partitioning, i.e., the input of each load. Branch circuit energized With power output The value of .

[0051] During the lower-level verification process, AC power flow calculation and frequency response verification are performed on the candidate islanding schemes obtained from the upper level to verify whether the bus voltage of each island meets the voltage threshold range, whether the power flow of each branch exceeds the limit, and whether the frequency offset of each island meets the constraints.

[0052] If the lower-level check finds that a certain island has a bus voltage exceeding the limit, branch power flow exceeding the limit, or frequency exceeding the limit, then a cut constraint is generated based on the limit information and sent back to the upper-level model. The cut constraint includes: excluding the branch combination that causes the limit to exceed the limit, or increasing the load reduction of the corresponding island. Then, the mixed integer linear programming model is re-executed to solve the problem until the generated island partitioning scheme passes the lower-level check or reaches the set maximum number of iterations.

[0053] In this embodiment, according to the grid islanding strategy, the converged branches are energized. With switch operation state Determine the boundary switches and their on / off states required for dividing each isolated island, and generate a sequence of switch operations.

[0054] In other embodiments, when multiple candidate islanding schemes are generated for different trigger times or different parameter settings, the evaluation indicators of each scheme are calculated separately. The evaluation indicators include at least the priority weighted recovery amount of critical loads, the total load loss, the degree of bus voltage deviation, the degree of system frequency deviation, and the number of switching operations.

[0055] At this point, the candidate schemes are comprehensively ranked according to the preset index weights, and the optimal scheme is selected. The comprehensive ranking takes the critical load recovery amount as the primary index. When the critical load recovery amounts are similar, the scheme with smaller frequency and voltage deviations and fewer switching operations is preferred.

[0056] In this embodiment, the grid islanding strategy planning could be periodically repeated as the typhoon wind field data is updated, thereby dynamically and continuously correcting the islanding scheme to adapt to the continuous changes in the grid damage situation during the typhoon's advance.

[0057] As a second embodiment of this application, a power grid islanding system based on topology reconfiguration and critical loads includes: The model building unit is used to build a power grid model based on power grid structure data, which includes topological connections, geospatial information and electrical operating parameters. The wind field assessment unit is used to spatially couple typhoon wind field data with the geospatial information of each component in the power grid model to obtain the wind field intensity experienced by each component at each time. The failure assessment unit is used to determine the failure risk of each component based on the wind field intensity according to the component damage assessment model. The topology reconfiguration unit is used to determine the faulty components that will be taken out of service based on the failure risk along the typhoon's progression, and to reconfigure the topology connections of the power grid model. The fault inference unit is used to perform power flow calculations on the reconstructed topology connections and obtain fault evolution inference results. The islanding unit pre-stores the islanding objective function, which is used to solve the islanding objective function based on the reconstructed topology connection relationship and output the grid islanding strategy when the fault evolution simulation result meets the grid islanding trigger condition.

[0058] The model building unit is used to construct a power grid model based on power grid structure data, including topological connections, geospatial information, and electrical operating parameters. Specifically, as the data foundation of the entire system, the model building unit is configured to read multi-source heterogeneous data from SCADA, GIS, and PMS systems and integrate them into a unified graph model data structure. Within this unit, not only are electrical connection matrices established between nodes (such as power node G1 and load node N1) and branches (such as transmission lines L1 and L2), but the geographic coordinate vector data of each component is also synchronously associated with electrical parameters such as resistance and reactance. This integrated modeling approach allows subsequent units to directly obtain complete component objects with both electrical and spatial attributes without repeatedly querying different databases, significantly improving the system's data access efficiency and real-time processing. It should be understood that the model building unit can also be configured with a data verification and cleaning submodule to automatically identify and correct abnormal data such as topological islands and missing parameters during the modeling process, ensuring the integrity and accuracy of the input data.

[0059] The wind field assessment unit spatially couples typhoon wind field data with the geospatial information of various components in the power grid model to obtain the wind field intensity experienced by each component at each time. Specifically, the wind field assessment unit accesses typhoon wind field raster data or numerical forecast products released by meteorological departments in real time through a communication interface, and uses a spatial indexing algorithm to quickly match it with the power grid geographic information output by the model building unit. For components with spatial extension attributes, such as transmission lines L1 and L2, this unit integrates a spatial intersection calculation engine, which can automatically calculate the representative wind speed when the line corridor crosses multiple wind speed grids (e.g., taking the maximum value or weighted average). To cope with the computational pressure of massive spatiotemporal data during typhoon passage, the wind field assessment unit can adopt a parallel computing architecture or GPU acceleration technology to ensure that the wind field intensity update of all network components is completed within a time step of minutes or even seconds, providing highly timely input data for subsequent failure risk assessment.

[0060] Specifically, the failure assessment unit incorporates a component damage assessment model. This unit receives the time-series wind field intensity sequence output by the wind field assessment unit and calculates the failure probability curves of each component over a future period. By embedding the physical mechanism model into the system's calculation process, the failure assessment unit achieves a quantitative mapping from meteorological elements to equipment health status, overcoming the limitations of traditional systems that rely solely on empirical thresholds, and making the risk assessment results more closely aligned with actual physical laws.

[0061] The topology reconfiguration unit maintains a dynamically evolving power grid graph model instance. At each simulation or prediction time, based on the failure risk value output by the failure assessment unit and pre-defined fault determination conditions, this unit automatically identifies and marks faulty components that have exited operation. Subsequently, the unit triggers a graph topology update operation, disconnecting the corresponding branch connections, modifying the adjacency matrix, and synchronously updating the status information of associated circuit breakers. This dynamic reconfiguration mechanism enables the system to realistically simulate the continuous changes in the power grid topology under typhoon disasters, rather than only processing static snapshots, thus providing a time-varying network environment that conforms to the evolution of disasters for fault simulation.

[0062] The fault simulation unit integrates a high-performance power flow calculation engine and connectivity analysis algorithms. Whenever the topology reconfiguration unit completes a network update, the fault simulation unit automatically initiates the analysis process: first, it identifies the current electrical island partitioning and source-load distribution using a graph search algorithm; then, it performs AC or DC power flow calculations on each electrical island. The calculation results are structured into fault evolution simulation results, including but not limited to a list of voltage-overriding buses, a list of power flow overload branches, a list of power-loss loads, and an estimated system frequency offset. These results not only visually represent the disaster situation but also serve as crucial state trigger signals transmitted to downstream units, achieving seamless integration from situational awareness to decision-making and response.

[0063] The islanding unit internally stores a mixed-integer linear programming model with the objective of maximizing load recovery. This model comprehensively considers the priority weights of critical loads, the cost of switching operations, and the penalty for power supply path failure risks. When it receives a trigger signal from the fault simulation unit (such as excessive load loss or backbone network disconnection), the unit automatically invokes the optimization solver to search for the optimal islanding scheme on the current damaged topology. Simultaneously, the islanding unit is equipped with a verification feedback mechanism that verifies the AC power flow and frequency stability of candidate schemes. If the verification fails, it automatically generates cut constraints and iteratively recalculates until a final strategy that satisfies both the optimization objective and physical feasibility is output. This two-layer architecture design, combining optimization and verification, effectively ensures the executability and security of the system's output scheme in actual engineering projects.

[0064] At the hardware implementation level, the aforementioned functional units can be deployed on the same high-performance server or distributed computing cluster. The system hardware includes at least a processor, memory, and a communication interface. The memory stores computer program instructions that implement the functions of each unit, as well as power grid model data, wind farm data, etc.; the processor is configured to execute these instructions to complete data processing and logical operations; the communication interface is used for data interaction with external meteorological systems and power grid dispatch automation systems. In terms of software architecture, each unit can adopt a microservice or modular design, decoupled and collaborative through standard API interfaces, facilitating subsequent functional expansion and maintenance upgrades. For example, when a new component vulnerability model needs to be introduced, only the algorithm library of the failure assessment unit needs to be updated without modifying other modules; when the power grid scale expands, the processing speed of wind farm assessment and fault inference can be improved by adding computing nodes in parallel. This hardware-software collaborative system design ensures the flexibility, scalability, and efficiency of the technical solution in engineering applications.

[0065] The specific embodiments described above are preferred embodiments of the power grid islanding method and system based on topology reconfiguration and critical loads in this application, and are not intended to limit the specific implementation scope of this application. The scope of this application includes but is not limited to the specific embodiments described above. All equivalent changes made in accordance with the shape and structure of this application are within the protection scope of this application.

Claims

1. A method for islanding power grids based on topology reconfiguration and critical loads, characterized in that: Includes the following steps: A power grid model is constructed based on power grid structure data, including topological connections, geospatial information, and electrical operating parameters. Acquire typhoon wind field data, and spatially couple the typhoon wind field data with the geospatial information of each component in the power grid model to obtain the wind field intensity experienced by each component at each time. The failure risk of each component is determined based on the intensity of the wind field effect, according to the component damage assessment model. Based on the failure risk advancing along the typhoon timeline, the faulty components that will be taken out of service are determined, and the topological connections of the power grid model are reconstructed. Perform power flow calculations on the reconstructed topology connections to obtain fault evolution prediction results; The degree of load recovery is calculated based on the critical load recovery amount and the penalty for power supply path failure risk. An islanding objective function is constructed with the goal of maximizing load recovery. When the fault evolution simulation results meet the grid islanding triggering conditions, the grid islanding strategy is output by solving the islanding objective function based on the reconstructed topology connection relationship.

2. The grid islanding method based on topology reconfiguration and critical loads as described in claim 1, characterized in that: The power grid model constructed based on power grid structure data, which includes topological connections, geospatial information, and electrical operating parameters, includes: The topological connection relationship is constructed by using power grid nodes as vertices of the graph model and power grid components as branches of the graph model. Electrical operating parameters are assigned to each branch, and geospatial information is assigned to each vertex and branch. Elements with spatial extension properties are geometricated into spatial line segments.

3. The grid islanding method based on topology reconfiguration and critical loads as described in claim 2, characterized in that: The spatial coupling of typhoon wind field data with the geospatial information of various components in the power grid model includes: At each moment, the typhoon wind field data is spatially intersected with the geospatial information of each vertex in the spatial line segment and power grid model, and the wind speed at the location of each component is extracted as the wind field intensity. For a spatial line segment that spans multiple spatial grid cells of typhoon wind field data, the maximum wind speed or the average wind speed weighted by element length in each spatial grid cell through which the spatial line segment passes is taken as the wind field intensity of the element corresponding to the spatial line segment.

4. The grid islanding method based on topology reconfiguration and critical loads as described in claim 1, characterized in that: The component damage assessment model uses the logistic function, with the critical wind resistance of the component as the inflection point parameter of the logistic function.

5. The grid islanding method based on topology reconfiguration and critical loads as described in claim 2, characterized in that: The process of determining the out-of-operation faulty components based on the failure risk along the typhoon's progression sequence and reconstructing the topology of the power grid model includes: At each moment of the typhoon propulsion sequence, when the risk of component failure meets the preset fault judgment conditions, the component is determined to be a faulty component that is taken out of operation. In response to the switch state changes caused by the exit of a faulty component, the connectivity between components is updated, the exited branches are deleted, and the topology of the power grid model is reconstructed.

6. The grid islanding method based on topology reconfiguration and critical loads as described in claim 1, characterized in that: The step of performing power flow calculations on the reconstructed topology connections to obtain fault evolution prediction results includes: The reconstructed topology is searched for connected components to identify electrical islands and their power and load distribution. For each electrical island, perform power flow calculations to obtain the voltage over-limit bus, power flow over-limit branch, and power failure load, which are used as the results of fault evolution prediction.

7. The grid islanding method based on topology reconfiguration and critical loads as described in claim 2, characterized in that: The calculation of load recovery level based on critical load recovery amount and power supply path failure risk penalty includes: The load recovery amount weight is obtained based on the key priority of the load, and the key load recovery amount is calculated using the load recovery amount weight and each load recovery amount. Calculate the power supply path failure risk penalty based on the branch failure risk; The degree of load recovery is calculated based on the difference between the critical load recovery amount and the penalty for power supply path failure risk.

8. The grid islanding method based on topology reconfiguration and critical loads as described in claim 7, characterized in that: The calculation of load recovery level based on critical load recovery amount and power supply path failure risk penalty also includes: Calculate the action loss based on the switching operation cost of the branch and the switching action state; The degree of load recovery is calculated by subtracting the action loss and the penalty for power supply path failure from the critical load recovery amount.

9. The grid islanding method based on topology reconfiguration and critical loads as described in claim 1, characterized in that: The grid islanding strategy, which outputs the objective function for islanding based on the reconstructed topology connectivity, includes: Based on the reconstructed topological connectivity, the objective function for island partitioning is solved, and candidate island partitioning schemes are output. Perform AC power flow calculations and frequency response verifications on the candidate islanding schemes; If the verification fails, cut constraints are generated based on the out-of-limit information. These cut constraints are then incorporated into the constraints of the island partitioning objective function. The island partitioning objective function is then solved again until the verification passes or the maximum number of iterations is reached.

10. A power grid islanding system based on topology reconfiguration and critical loads, used to implement the method as described in any one of claims 1 to 9, characterized in that: include: The model building unit is used to build a power grid model based on power grid structure data, which includes topological connections, geospatial information and electrical operating parameters. The wind field assessment unit is used to spatially couple typhoon wind field data with the geospatial information of each component in the power grid model to obtain the wind field intensity experienced by each component at each time. The failure assessment unit is used to determine the failure risk of each component based on the wind field intensity according to the component damage assessment model. The topology reconfiguration unit is used to determine the faulty components that will be taken out of service based on the failure risk along the typhoon's progression, and to reconfigure the topology connections of the power grid model. The fault inference unit is used to perform power flow calculations on the reconstructed topology connections and obtain fault evolution inference results. The islanding unit pre-stores the islanding objective function, which is used to solve the islanding objective function based on the reconstructed topology connection relationship and output the grid islanding strategy when the fault evolution simulation result meets the grid islanding trigger condition.

Images