Source network load storage collaborative active defense regulation method and system, computer device and computer program product

Abstract

The application belongs to the technical field of power grid safety prevention and control, and particularly relates to a source-grid-load-storage collaborative active defense regulation method and system, computer equipment and computer program products, through hierarchical cooperation of a city group level coordination center, a city level dispatching center and a station level execution terminal, a typhoon whole-process closed-loop decision is realized: in a prevention stage, an IGDT robust optimization is used to generate a pre-dispatching scheme resistant to prediction errors, to reserve adjustment margin for subsequent stages; in a response stage, MPC rolling emergency control and dynamic hosting of failed areas are used to minimize real-time losses; in a recovery stage, preset energy storage black start resources are used to accelerate power restoration. The three-stage instructions are executed by the station level terminal and feedback is formed from decision making to execution to feedback closed loop. The method significantly reduces the risk of large-area power outages in a city group, shortens the recovery time and improves the robustness to typhoon uncertainty.

Description

Technical Field

[0001] This application belongs to the field of power grid security and control technology, and in particular relates to a source-grid-load-storage coordinated active defense and control method, system, computer equipment and computer program product. Background Technology

[0002] Typhoons are among the most frequent and destructive extreme weather events threatening the power grid security of coastal urban agglomerations. In recent years, with the deepening of the construction of new power systems, the flexible adjustment resources of all aspects of power generation, grid, load, and storage have become increasingly abundant, providing a new material foundation for enhancing the disaster resilience of the power grid. At the same time, the electrical connections between the power grids of various cities within urban agglomerations are becoming increasingly close, and cross-regional power support and reserve sharing have become routine operational requirements. However, in typhoon disaster scenarios, traditional decision-making models still rely primarily on each city independently formulating emergency plans, resulting in limited information sharing and a lack of coordination mechanisms, making it difficult to leverage the overall resource advantages of the urban agglomeration. Forecast errors in typhoon paths and intensity always exist, significantly reducing the effectiveness of offline plans based on deterministic scenarios in practical applications. Furthermore, power grid dispatching and operation inherently possess the characteristics of a "hierarchical and regionalized, locally managed" system, and centralized global optimization decision-making faces obstacles in terms of data privacy, communication reliability, and the division of responsibilities. Against this backdrop, there is an urgent need to construct a coordinated decision-making methodology that respects the existing management system, achieves efficient collaboration at the urban agglomeration level, covers the entire typhoon process, and robustly addresses forecast uncertainties.

[0003] Existing technologies typically focus on offline quantitative coordination in a single region. Their advantages lie in their ability to better solve single-dimensional or local problems: centralized plans can theoretically achieve global optimality, distributed coordination can protect the data privacy of each city and be compatible with the existing system, rolling emergency control can respond quickly to real-time disturbances and effectively handle temporal coupling constraints, and robust optimization can provide a clear measure for decision-making to resist forecast errors.

[0004] However, all of these methods have fundamental limitations: centralized contingency plans cannot dynamically respond to real-time changes in typhoons and conflict with hierarchical management systems; distributed coordination lacks a unified risk quantification benchmark for disaster scenarios and is limited to normal operating conditions; rolling emergency control is not effectively coupled with cross-regional coordination mechanisms; and robust optimization is mostly a one-time pre-disaster decision for a single region, failing to form a full-cycle closed loop and is difficult to implement in a distributed architecture. These methods are isolated from each other and cannot simultaneously meet the comprehensive needs of urban power grids under extreme typhoons for unified risk measurement, full-time connectivity, uncertainty robustness, distributed feasibility, self-governance in failure, and multiple objective options. This leads to coordination decisions in extreme scenarios often falling into the dilemma of "local optima, overall failure." Summary of the Invention

[0005] To address the technical problems mentioned in the background art regarding the fragmented decision-making and insufficient coordination of urban power grids under typhoon conditions, this application provides a source-grid-load-storage coordinated active defense and control method, system, computer equipment, and computer program product.

[0006] In a first aspect, embodiments of this application provide a source-grid-load-storage coordinated active defense and control method. The method is applied to a source-grid-load-storage coordinated active defense and control system, which includes a city cluster-level coordination center, a city-level dispatch center, and station-level execution terminals. The method includes: Step S1: The city cluster-level coordination center acquires the operation status data, equipment vulnerability parameters, and typhoon forecast scenario sets of the power grids of each city within the city cluster; the city cluster-level coordination center uses the operation status data, equipment vulnerability parameters, and typhoon forecast scenario sets belonging to different risk types in the power grids of each city as heterogeneous risk elements, uniformly maps them to the expected load loss risk value RaV of each city, and constructs the typhoon spatiotemporal stochastic evolution scenario set and its probability distribution uncertainty set; the city cluster-level coordination center also initializes the alternating direction multiplier method (ADMM) communication parameters between itself and the city-level dispatch center. Step S2: The city cluster-level coordination center automatically identifies the current stage type based on the time tag and geofence information in the typhoon spatiotemporal random evolution scenario set. The stage type includes prevention stage, response stage, and recovery stage. The city cluster-level coordination center uniformly issues stage switching instructions to each city-level dispatch center. Step S3: When the phase switching instruction received by the city-level dispatch center is the prevention phase, the city-level dispatch center constructs a multi-objective Pareto optimization model based on the expected load loss risk value RaV, with the four-dimensional objectives of minimizing the total expected load loss of the system, minimizing the maximum risk range of the system, minimizing the total economic cost, and minimizing the recovery time; the city-level dispatch center uses the information gap decision theory (IGDT) to set the key typhoon parameters in the typhoon forecast scenario set as uncertain variables, and solves for a robust prevention dispatch scheme that can withstand the maximum forecast error; the city-level dispatch center... The city-level dispatch center uses the ADMM distributed iterative solution initialized with the ADMM communication parameters to solve the local optimization subproblems in parallel, gradually converging to the global Pareto optimal solution, while only exchanging the expected power values ​​of the tie lines with the city cluster-level coordination center. The city-level dispatch center ultimately generates preventative control commands, including topology reconfiguration, forced full charging of energy storage, unit hot standby switching, and load peak shifting. The city-level dispatch center then issues these preventative control commands to the station-level execution terminals for execution. These preventative control commands reserve adjustment margins for subsequent response phases. Step S4: When the phase switching instruction received by the city-level dispatch center is in the response phase, the city-level dispatch center, based on the energy storage forced full charge state and unit hot standby switching state configured in the prevention and control instruction, collects wide-area measurement signals and fault signals in real time at fixed time periods to detect the communication and functional status of each area; for the detected failure areas, the city cluster-level coordination center automatically dynamically divides the failure areas according to the electrical adjacency matrix and entrusts them to the city-level dispatch centers of adjacent healthy cities; the city-level dispatch center adopts the model predictive control (MPC) method to minimize the load loss, frequency deviation, and voltage deviation in the current and predicted time domains, and solves the emergency control instructions in a rolling manner. The emergency control instructions include new energy power reduction flight, energy storage network support, distribution network island supply guarantee, and automatic load shedding; the city-level dispatch center issues the emergency control instructions to the station-level execution terminal for execution, and only executes the emergency control instructions of the current time step, and repeats the rolling optimization in the next cycle; the city-level dispatch center uses the fault records and system state changes in the response phase as the initial conditions for the recovery phase; Step S5: When the phase switching instruction received by the city-level dispatch center is the recovery phase, the city-level dispatch center, based on the system state after the response phase, statistically analyzes the equipment outage status, remaining energy storage capacity, and available power generation capacity, and constructs a black-start recovery optimization model with the goal of minimizing the maximum recovery time of the entire system; the city-level dispatch center prioritizes calling the energy storage system that has been forcibly fully charged in the prevention and control instruction and whose energy was not depleted in the response phase as the black-start power source, and generates a black-start recovery instruction; the city-level dispatch center issues the black-start recovery instruction to the station-level execution terminal for execution, establishes an islanded microgrid, and restores distributed power sources and distribution network lines in sequence; the city-level dispatch center also solves the cross-regional power support scheme through ADMM coordination between the city cluster-level coordination center and healthy cities; the city-level dispatch center puts in cold loads in batches according to the impact coefficient and importance level, and after each batch is put in, it verifies the voltage and frequency stability through real-time data feedback from the station-level execution terminal; Step S6: Throughout the typhoon event, the city cluster-level coordination center continuously monitors its communication status with each city-level dispatch center. When the city cluster-level coordination center is detected to be paralyzed, each city-level dispatch center automatically switches to its local pre-set conservative strategy to maintain minimum safe operation. After communication is restored, the state is smoothly synchronized and ADMM coordination is restarted. The city-level dispatch center also displays a probabilistic auxiliary decision panel to the dispatcher, which includes the optimal solution, success probability, robust radius, and list of weak nodes. It responds to the dispatcher's online adjustment instructions for optimization parameters and triggers re-optimization. After the typhoon passes, the city cluster-level coordination center performs post-event settlement of cross-regional support based on the ADMM multiplier trajectory and feeds back the model parameters and effects of this event to the historical database.

[0007] The technical advantages of this application are as follows: Through hierarchical collaboration between the city cluster-level coordination center, the city-level dispatch center, and the station-level execution terminal, a closed-loop decision-making process for the entire typhoon process is achieved: In the prevention phase, robust optimization using IGDT generates a pre-dispatch scheme that can withstand forecast errors, reserving adjustment margins for subsequent phases; in the response phase, rolling emergency control using MPC and dynamic management of failed areas minimizes real-time losses; and in the recovery phase, pre-set energy storage black-start resources accelerate power restoration. All three phases of instructions are executed and fed back by the station-level terminal, forming a "decision-execution-feedback" closed loop. This method significantly reduces the risk of large-scale power outages in city clusters, shortens recovery time by 30%-50%, improves robustness to typhoon uncertainties, and the ADMM distributed architecture protects the data privacy of each city, is compatible with the existing dispatch system, and has high engineering practicality.

[0008] In a specific implementation of the first aspect, step S1, constructing a set of spatiotemporal stochastic evolution scenarios of typhoons and its probability distribution uncertainty set, includes: The city cluster-level coordination center constructs a set of spatiotemporal random evolution scenarios for typhoons. Each scenario in the set includes temporal wind speed, spatial distribution of rainfall, and probability weights. The city cluster-level coordination center uses φ-divergence to construct a probability distribution uncertainty set:

[0009] Where κ is the divergence threshold, which represents the allowable deviation between the actual probability distribution and the nominal probability distribution.

[0010] The technical effect of this application's embodiments is that by constructing an uncertainty set using φ-divergence, compared to traditional deterministic scenarios or simple probabilistic scenarios, it can more compactly characterize the statistical deviation range of typhoon path and intensity, providing a mathematically rigorous uncertainty measurement benchmark for subsequent robust IGDT optimization, and avoiding decision failure due to probability distribution estimation bias.

[0011] In a specific implementation of the first aspect, in step S3, the multi-objective Pareto optimization model uses the enhanced ε-constraint method to transform the other three objectives, except for the main objective of minimizing the total expected load loss of the system, into constraints. By changing the ε value, a complete Pareto front solution set is generated so that the scheduler can select the corresponding scheme according to actual preferences.

[0012] The technical effect of this application's embodiments: The enhanced ε-constraint method can generate a uniformly distributed Pareto front in the four-dimensional target space, avoiding the shortcomings of the traditional weighted sum method in handling non-convex fronts, enabling decision-makers to intuitively see the objective trade-off between reliability, safety, economy and recovery speed, and achieving transparent selection.

[0013] In one specific implementation of the first aspect, step S4, the model predictive control (MPC) method includes: At each sampling time k, solve the following rolling optimization problem:

[0014] in, Let be the load shedding amount at time t. For frequency deviation, For critical bus voltage deviation; These are the weighting coefficients. The unit penalty coefficient for frequency deviation, This is the unit penalty factor for voltage deviation; The constraints of the rolling optimization problem include: System frequency response swing equation constraints: Where H is the system inertia constant and D is the load damping coefficient; Let be the change in mechanical power at time t. Let be the change in electromagnetic power at time t; Constraints on the dynamic equations of the energy storage state of charge (SOC): ,in, For charge and discharge efficiency; Let be the remaining charge of energy storage b at time t.

[0015] In a specific implementation of the first aspect, in step S5, the objective function of the black start recovery optimization model is:

[0016] in, A collection of energy storage devices with black start capability. Let be the recovery time for the b-th energy storage unit.

[0017] The technical effects of this application's embodiments are as follows: by adopting a minimization-maximization objective function, the recovery time of the slowest energy storage is directly optimized, avoiding delays in the overall system's power restoration progress due to the lag in the recovery of a single energy storage; the charging reactive power limitation and voltage drop constraint jointly prevent secondary accidents such as overvoltage or voltage collapse during the recovery process, ensuring the safety of the recovery process itself.

[0018] In a specific implementation of the first aspect, in step S3, the convergence criterion for the ADMM distributed iterative solution is: Calculate the original residuals and dual residual ; in the original residual And dual residuals In the case of convergence, it is determined to be convergent. in, For tie-line power variables, It is a copy of its global variable, and k is the iteration round.

[0019] The technical effects of the embodiments of this application are as follows: the dual residual convergence criterion takes into account both the original feasibility (the power of the tie line is consistent among cities) and the dual feasibility (the multiplier update is stable), which can reliably terminate the iteration and avoid invalid calculations; the criterion directly corresponds to the convergence proof of ADMM theory, ensuring that the deviation between the distributed solution results and the centralized global optimal solution is controllable.

[0020] In one specific implementation of the first aspect, the dynamic partitioning in step S4 is specifically implemented as follows: The city cluster-level coordination center continuously monitors the heartbeat signals of each city-level dispatch center. When the heartbeat signal times out or the measurement freezes, indicating a loss of communication or function in the affected area, the city cluster-level coordination center determines the cause based on the electrical adjacency matrix. The lost-connection area is automatically assigned to the city-level dispatch center of a neighboring healthy city. The city-level dispatch center of the healthy city incorporates a simplified equivalent model of the lost-connection area into its local optimization subproblem and uses residual measurements as pseudo-boundary conditions for unified optimization.

[0021] The technical effects of this application's embodiments: The dynamic partitioning and hosting mechanism enable the method to survive under extreme conditions: When some areas lose their scheduling function due to a typhoon, adjacent healthy cities can automatically take over and control the situation, avoiding the paralysis of the entire coordination framework due to a single point of communication interruption, and greatly improving the robustness and availability of the decision-making system in real extreme disaster environments.

[0022] Secondly, this application also proposes a source-grid-load-storage coordinated active defense and control system, which includes a city cluster-level coordination center, a city-level dispatch center, and a station-level execution terminal; the city cluster-level coordination center, the city-level dispatch center, and the station-level execution terminal cooperate with each other to execute the steps of the method described in the first aspect above.

[0023] Thirdly, this application also proposes a computer device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the method described in the first aspect above.

[0024] Fourthly, this application also proposes a computer program product comprising a computer program that, when executed by a computer device, implements the steps of the method described in the first aspect above.

[0025] For the technical effects of the second to fourth aspects of the invention mentioned above, please refer to the relevant technical effects of the first aspect, which will not be repeated here. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the embodiments of this application, 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 of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0027] Figure 1 A schematic diagram of the main process of a source-grid-load-storage coordinated active defense and control method provided in this application; Figure 2 A flowchart illustrating an embodiment of the source-grid-load-storage coordinated active defense and control method provided in this application; Figure 3 A structural block diagram of a source-grid-load-storage coordinated active defense and control device provided in this application; Figure 4 A schematic diagram of an embodiment of the computer device for executing the source-grid-load-storage collaborative active defense and control method provided in this application. Detailed Implementation

[0028] 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 application pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the drawings are intended to cover non-exclusive inclusion.

[0029] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0030] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0031] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if detected [the described condition or event]" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once detected [the described condition or event]," or "in response to detection [the described condition or event]."

[0032] In the description of the embodiments of this application, the term "multiple" refers to two or more (including two), unless otherwise expressly and specifically defined.

[0033] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0034] References to "one embodiment" or "some embodiments" in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized.

[0035] The terms “including,” “comprising,” “having,” and variations thereof all mean “including but not limited to,” unless otherwise specifically emphasized.

[0036] To address the technical problems mentioned in the background section, firstly, referring to... Figure 1 This application provides a source-grid-load-storage coordinated active defense control method, which is applied to a source-grid-load-storage coordinated active defense control system. (Refer to...) Figure 3 The source-grid-load-storage collaborative active defense and control system mainly includes a city cluster-level coordination center, a city-level dispatch center, and station-level execution terminals; It should be noted that the city cluster-level coordination center in this application embodiment can be a server (one or more) deployed on the upper-level dispatching agency; the city-level dispatching center in this application embodiment can be a server deployed on the dispatching agencies of each city (one set for each city); the station-level execution terminal in this application embodiment can be an edge intelligent terminal (such as PLC, RTU, protection device) deployed on the field of substations, power plants, energy storage stations, etc.

[0037] Accordingly, the source-grid-load-storage coordinated active defense and control method of this application includes the following S1-S6: Step S1 (Risk Modeling and Initialization): The city cluster-level coordination center acquires the operational status data, equipment vulnerability parameters, and typhoon forecast scenario sets of the power grids of each city within the city cluster; the city cluster-level coordination center uses the operational status data, equipment vulnerability parameters, and typhoon forecast scenario sets belonging to different risk types in the power grids of each city as heterogeneous risk elements, uniformly maps them to the expected load loss risk value RaV of each city, and constructs the spatiotemporal stochastic evolution scenario set of typhoons and its probability distribution uncertainty set; the city cluster-level coordination center also initializes the alternating direction multiplier method (ADMM) communication parameters between itself and the city-level dispatch center. As a specific implementation, after the typhoon forecast information is released, this embodiment first performs risk modeling and initialization steps, which further includes sub-steps 1.1 to 1.4: Sub-step 1.1 Data Acquisition: The city cluster-level coordination center acquires the following three types of data from the energy management system (EMS) and meteorological information system of each city-level dispatch center through the dispatch data network: The first category is operational status data, including real-time load curves of each city's power grid, planned and actual output of generator units, state of charge (SOC) of energy storage stations, current switching power of tie lines, power flow of lines and transformers, and bus voltage. Taking city A as an example, the current total load is 3200MW, the available generating capacity is 2800MW, the total energy storage capacity is 200MWh / 100MW, and the current transmission power of the AB tie line is 150MW (A to B). The second category consists of equipment vulnerability parameters, including failure probability curves for equipment such as transmission lines, transformers, and towers under different wind speed ranges. For example, the failure probability of a 220kV line is 0.2 at a wind speed of 25m / s, 0.7 at 35m / s, and 0.95 at 45m / s; wind turbines begin to trip protection when the wind speed exceeds 25m / s, etc. The third category is the typhoon forecast scenario set, which contains time-series forecast data for S typhoon paths. In this embodiment, the city cluster-level coordination center obtains 100 random simulated typhoon paths (S=100) from the meteorological department. Each path contains information such as the typhoon center location, maximum wind speed, wind circle radius, and rainfall intensity for each hour of the next 72 hours. Each path corresponds to an occurrence probability p_s^0.

[0038] Sub-step 1.2: Heterogeneous risk factors are uniformly mapped to... RaV The city cluster-level coordination center treats the three types of data—which have different sources, different dimensions, and different properties—as heterogeneous risk factors and uniformly maps them to the expected load loss risk value for each city. RaV The mapping formula is as follows:

[0039] in, i The city identifiers are represented by i (i=A,B,C), and s represents the typhoon scene. Let be the probability of scenario s occurring. For the city in scene s i The predicted load shedding amount, For the city in this scenario The load shedding amount. This enables symmetrical comparability of risk states across multiple cities and voltage levels; For example, The prediction is based on the following calculations: First, based on the wind speed at each moment in scenario s and combined with the equipment vulnerability parameters, the Monte Carlo method is used to sample and determine the fault state of each device; then, based on the network topology after the fault, the minimum load shedding amount is calculated using the optimal load shedding model. In this embodiment, it was calculated that: City A RaV =125MWh, City B RaV =80MWh, City C RaV =210MWh; RaV The higher the value, the higher the expected risk of loss for the city during a typhoon.

[0040] Sub-step 1.3 Constructing the Spatiotemporal Stochastic Evolution Scenarios and Uncertainty Set for Typhoons: The city cluster-level coordination center constructs a spatiotemporal stochastic evolution scenario set for typhoons, where each scenario includes temporal wind speed, spatial distribution of rainfall, and probability weights. To handle the uncertainty in typhoon forecasts, the city cluster-level coordination center uses φ-divergence to construct a probability distribution uncertainty set:

[0041] Wherein, κ is the divergence threshold. In this embodiment, κ=0.5, which means that a certain deviation is allowed between the actual probability distribution and the nominal probability distribution, so as to reflect the inherent error of typhoon track and intensity forecast.

[0042] Sub-step 1.4 Initialize ADMM communication parameters: The city cluster-level coordination center initializes its own ADMM communication parameters with each city-level dispatch center, including: Initialization of global variables (Indicates that the expected power of the link between city i and city j is initially 0), initial values ​​of Lagrange multipliers. Penalty parameter ρ > 0, original residual convergence threshold =0.01MW, dual residual convergence threshold =0.01MW.

[0043] At this point, the initialization work in step S1 is complete, and the system enters the scheduling state.

[0044] Step S2 (Phase Determination): The city cluster-level coordination center automatically identifies the phase type of the current time period based on the time tag and geofence information in the typhoon spatiotemporal random evolution scenario set. The phase type includes prevention phase, response phase and recovery phase. The city cluster-level coordination center uniformly issues phase switching instructions to each city-level dispatch center. It is understandable that step S2 involves the city cluster-level coordination center automatically identifying the current stage type based on the time tags and geofence information in the typhoon spatiotemporal random evolution scenario set constructed in step S1. This step can further include the following sub-steps: Sub-step 2.1 Time stamp determination: The city cluster-level coordination center reads the current system time T. now And the typhoon's expected landfall time T landfall In this embodiment, the current time is 48 hours before the typhoon is expected to make landfall; it can be stipulated that: the prevention phase corresponds to 24-72 hours before landfall, the response phase corresponds to the direct impact period (approximately 6-12 hours before and after landfall), and the recovery phase corresponds to after the typhoon has passed.

[0045] Sub-step 2.2 Geofencing Determination: The city cluster-level coordination center also makes auxiliary determinations based on the relative relationship between the typhoon's real-time location and the geographical scope of the city cluster. When the typhoon center enters within 200 kilometers of the coastline but has not yet made landfall, it is marked as the late stage of the prevention phase; when the typhoon center enters land and the wind speed remains above 25 m / s, it is marked as the response phase; when the typhoon center moves out of the city cluster and the wind speed remains below 15 m / s, it is marked as the recovery phase.

[0046] Sub-step 2.3 Issuing a phase switching instruction: In this embodiment, the current time is 48 hours before the landing, which is determined to be the prevention phase; the city cluster coordination center uniformly issues a phase switching instruction to the city-level dispatch centers of cities A, B, and C. The instruction content is "current phase = prevention phase, estimated duration = 48 hours", and carries the RaV value of each city for subsequent optimization. If the typhoon's path changes, the city cluster-level coordination center will update its assessment in real time and issue new phase switching instructions. For example, after the typhoon makes landfall, the city cluster-level coordination center will automatically update the phase switching instructions to "current phase = response phase" and distribute them to the dispatch centers of each city.

[0047] Step S3 (Multi-objective Robust Pre-scheduling and Cross-regional Coordination in the Prevention Phase): When the phase switching instruction received by the city-level dispatch center is for the prevention phase, the city-level dispatch center constructs a multi-objective Pareto optimization model based on the expected load loss risk value RaV, with four-dimensional objectives: minimizing the total expected load loss of the system, minimizing the maximum risk range of the system, minimizing the total economic cost, and minimizing the recovery time. The city-level dispatch center uses the information gap decision theory (IGDT) to set the key typhoon parameters in the typhoon forecast scenario set as uncertain variables, and solves for the robust pre-scheduling model that can withstand the maximum forecast error. The anti-dispatch scheme involves the city-level dispatch center using the ADMM distributed iterative solution initialized with ADMM communication parameters. Under the condition of exchanging tie-line power expectations only with the city-group-level coordination center, it solves local optimization subproblems in parallel and gradually converges to the global Pareto optimal solution. The city-level dispatch center ultimately generates preventative control commands, including topology reconfiguration, forced full charging of energy storage, unit hot standby switching, and load peak avoidance. The city-level dispatch center then issues these preventative control commands to the station-level execution terminals for execution. These preventative control commands are used to reserve adjustment margins for subsequent response phases. In this embodiment of the application, after the city-level dispatch centers of cities A, B, and C receive the phase switching instruction for the prevention phase, they execute the following sub-steps in parallel: Sub-step 3.1 Constructing a multi-objective Pareto optimization model: Based on the RaV value of their respective city calculated in step S1, each city-level dispatch center constructs a four-dimensional objective optimization model. Taking city A as an example, the optimization objective is: Objective f1 (reliability): Minimize the total expected load loss of the system, which is the probabilistic weighted sum of the load shedding amounts under all typhoon scenarios; Objective f2 (safety): Minimize the RaV range of each city, i.e., max(RaV_A,RaV_B,RaV_C) - min(RaV_A,RaV_B,RaV_C), to promote risk balance among cities; Objective f3 (economic efficiency): Minimize total operating costs, including unit fuel costs, energy storage charging and discharging losses, and load interruption compensation costs; Target f4 (Recovery Speed): Minimize the estimated maximum recovery time, i.e. the time required from the passage of the typhoon to the recovery of 90% of the load.

[0048] Since the four objectives conflict with each other (for example, improving reliability often requires increasing economic costs), this embodiment uses the enhanced ε-constraint method for solving the problem. Specifically, f1 is set as the primary objective, and f2, f3, and f4 are transformed into constraints f2 ≤ ε2, f3 ≤ ε3, and f4 ≤ ε4. By changing the values ​​of ε2, ε3, and ε4, a series of subproblems are generated. Solving each subproblem yields a set of Pareto optimal solutions, and the solutions to all subproblems constitute the Pareto front.

[0049] In this embodiment, after generating the Pareto front, the city-level dispatch center presents the front solution set to the dispatcher, who then selects the scheme ε2=50MWh, ε3=2 million yuan, and ε4=6 hours based on actual preferences (such as the current requirement for extremely high power supply reliability).

[0050] Sub-step 3.2 Robust Decision Modeling using IGDT: Considering the inherent errors in typhoon forecasting, this embodiment introduces the Information Gap Decision Theory (IGDT) to handle uncertainties. The maximum typhoon wind speed v is used as the basis for modeling. max As an uncertain parameter, its nominal value =40m / s (from weather forecast), where the radius of uncertainty β represents the maximum deviation of the actual wind speed from the nominal value.

[0051] The city-level dispatch center first sets an acceptable loss threshold L. c In this embodiment, the dispatcher in city A sets L. c =200MWh, meaning the maximum allowable expected load loss is no more than 200MWh.

[0052] Then, the maximum radius of uncertainty that can be tolerated is calculated:

[0053] in, It is a set of uncertainties; The robust prevention scheduling scheme ultimately selected is:

[0054] In this embodiment, the optimal solution x is obtained by solving. Corresponding maximum tolerance radius =0.25, meaning that even if the actual typhoon wind speed fluctuates between 30m / s and 50m / s (a deviation of ±25% from the forecast value of 40m / s), the expected load loss of the system will still not exceed 200MWh. This gives dispatchers clear confidence for decision-making.

[0055] Sub-step 3.3 ADMM Distributed Iterative Solution: To achieve cross-regional coordination while protecting the data privacy of each city, each city-level dispatch center uses the ADMM algorithm for distributed iterative solution. The mathematical form of the optimization problem is:

[0056] in, These are the internal decision variables for city i (unit start-up and shutdown, energy storage charging and discharging plans, etc.). The power of the connecting line between cities i and j. For the collection of connecting lines, This is the cost function related to the tie line.

[0057] The constraints include: Local execution constraints (Power balance, unit ramp-up, energy storage SOC, etc.), tie-line capacity constraints (Under the typhoon) , It is understood that capacity may be reduced under typhoon conditions; this embodiment is set accordingly. =200MW×(1 0.3) = 140MW (due to a 30% reduction in transmission capacity caused by the typhoon); Standby shared pool constraints That is, the total upper reserve is not less than the total lower reserve; and consistency constraints That is, the power at both ends of the tie line is an opposite of each other.

[0058] It is understandable that the core idea of ​​solving the ADMM in this embodiment is to relax the global consistency constraint into an augmented Lagrangian function, and each city independently solves its local subproblem; the augmented Lagrangian subproblem for city i is:

[0059] in, Let i be the set of cities adjacent to city i. This is a copy of the global variable for tie-line power.

[0060] The iterative process is as follows: No. Round of iteration: The city-level dispatch centers of cities A, B, and C solve their respective local optimization subproblems in parallel, obtaining... Only the expected value of the tie line power The data is reported to the city cluster-level coordination center without disclosing any internal operational data; subsequently, the coordination center updates the global variables.

[0061] Multiplier Update:

[0062] Convergence criterion: Calculate the original residuals and dual residual ; in the original residual And dual residuals In this case, convergence is determined. In this embodiment, ,and The convergence was determined under certain conditions, specifically after 15 iterations.

[0063] Sub-step 3.4 Generate and issue prevention and control instructions: After convergence, each city implements its own optimization scheme, and the tie-line power is adjusted according to... implement; Specifically, after convergence, each city-level dispatch center obtains the city's optimization plan, generates specific prevention and control instructions, and distributes them to the station-level execution terminals through the dispatch data network.

[0064] The prevention and control directives generated by City A include: (1) Topology reconfiguration instruction: Close the loop on two 10kV feeders in the northwest area of ​​City A, switch the open-loop structure under normal operation to a ring network structure, so as to enhance the power supply reliability of the area; at the same time, pre-decommission three high-risk 35kV lines (L102, L105, L108) located on the typhoon path to avoid line faults causing cascading trips during the typhoon.

[0065] (2) Forced full charge instruction for energy storage: Three energy storage power stations in City A (total capacity 200MWh) will be forced to charge from the current SOC=40% to SOC=100%, with the charging power being the rated power of each energy storage unit, and the process is expected to be completed in 6 hours. This instruction reserves sufficient emergency backup power for the subsequent response phase.

[0066] (3) Unit hot standby switching command: The two 300MW gas turbine units in City A switch from economic operation mode to hot standby mode, maintaining synchronous speed and grid connection but without load, and can respond to load changes within 5 minutes.

[0067] (4) Load shifting instruction: Through the demand response platform, a shifting production incentive notice was issued to 12 large industrial users to guide them to shift their production load to other time periods within 24 hours before the typhoon lands. It is expected to reduce the base load by about 80MW.

[0068] The station-level execution terminal receives and executes the above instructions. For example, after receiving the "forced full charge" instruction, the edge terminal of an energy storage power station starts the energy storage converter PCS, begins charging at a power of 500kW, and transmits the SOC data back to the city-level dispatch center in real time.

[0069] Step S4 (Rolling Emergency Control and Dynamic Zoning in the Response Phase): When the phase switching instruction received by the city-level dispatch center is in the response phase, the city-level dispatch center, based on the energy storage forced full-charge state and unit hot standby conversion state configured in the prevention and control instructions, collects wide-area measurement signals and fault signals in real time at fixed time periods to detect the communication and functional status of each area; for the detected failure areas, the city cluster-level coordination center automatically dynamically divides the failure areas according to the electrical adjacency matrix and entrusts them to the city-level dispatch centers of adjacent healthy cities; the city-level dispatch center adopts the Model Predictive Control (MPC) method to minimize the load loss, frequency deviation, and voltage deviation in the current and predicted time domains, and solves the emergency control instructions on a rolling basis. The emergency control instructions include new energy power reduction flight, energy storage network support, distribution network island supply guarantee, and automatic load shedding; the city-level dispatch center issues the emergency control instructions to the station-level execution terminals for execution, and only executes the emergency control instructions of the current time step, and repeats the rolling optimization in the next cycle; the city-level dispatch center uses the fault records and system state changes in the response phase as the initial conditions for the recovery phase; In its implementation, this step further includes the following sub-steps: Sub-step 4.1 Real-time monitoring and failure detection: Each city-level dispatch center continuously collects wide-area measurement signals (PMU data), real-time weather conditions (wind speed, rainfall), and fault signals (protection action information, circuit breaker position changes) every 15 minutes. Simultaneously, the city cluster-level coordination center continuously monitors the heartbeat signals of each city-level dispatch center—sending a "live" status packet every 5 seconds.

[0070] In this embodiment, one hour after the typhoon made landfall, the building where the dispatch center of City C is located experienced a power outage and communication link failure due to strong winds. The heartbeat signal timed out for more than 30 seconds, and the city cluster-level coordination center determined that the dispatch area of ​​City C had "loss of communication / function".

[0071] Sub-step 4.2 Failure area trusteeship transfer, the city cluster-level coordination center according to the electrical adjacency matrix Automatic dynamic partitioning is performed. As an example, in the electrical adjacency matrix, A AB =1 indicates that A and B are directly connected by a connecting line, A BC =1 indicates that B and C are directly connected by a connecting line, A AC =0 indicates that A and C have no direct connection; Because city C lost contact, the city cluster coordination center reassigned city C to city B (A), which has the closest electrical connection with it. BC =1), the city-level dispatch center of city B will temporarily "manage" the dispatch responsibilities of region C. The management operation includes: The B city dispatch center incorporated a simplified equivalent model of the C disconnected area into its local optimization subproblem. This model, based on the pre-typhoon topology and load characteristics of C city, used residual measurements in C area (such as voltage and power at substations on the C area boundary) as pseudo-boundary conditions and incorporated them into a unified optimization framework. The B city dispatch center formulated temporary emergency control strategies for C area and distributed them to the station-level terminals in C area through remaining available communication channels (such as backup fiber optic cables between B and C).

[0072] Sub-step 4.3 Model Predictive Control (MPC) Emergency Rolling Optimization: For cities A and B that still have scheduling capabilities, their city-level dispatch centers adopt the Model Predictive Control (MPC) method for rolling emergency control. As an example, the MPC configuration is as follows: The city-level dispatch center uses Δt = 15 minutes as a preset time step and constructs a rolling optimization problem with T = 4 preset time steps in the prediction time domain. At each sampling time k, the following optimization problem is solved:

[0073] in, Let be the load shedding amount at time t. For frequency deviation, For the critical bus voltage deviation; the weighting coefficient is taken as follows: =1000 yuan / MWh =500 yuan / Hz² =300 yuan / pu², reflecting the priority of load loss cost in emergency situations, followed by frequency stability, and then voltage stability.

[0074] Constraints can include: System frequency response swing equation: Where H is the system inertia constant (as an example, in this embodiment, H=5.2s for the city A system), and D is the load damping coefficient (taken as D=1.0).

[0075] Energy storage SOC dynamic equation: Where η is the charge / discharge efficiency (take η=0.92). It is worth noting that since the energy storage has been forcibly fully charged in step S3, the current SOC=100%, providing the maximum available discharge capacity for emergency control.

[0076] Step 4.4 Emergency Coordination Control of Source-Grid-Load-Storage: Based on the MPC optimization results, each city-level dispatch center generates emergency control commands and issues them to the station-level execution terminals for execution. (1) Reduced power flight of new energy sources: The coastal wind farm in City A (e.g., 300MW installed capacity) implements a reduced power flight strategy based on an instantaneous wind speed of 45m / s. The instruction is as follows: As an example, =180MW (current output) =10 minutes, meaning the output will be linearly reduced to 0 within 10 minutes to protect the wind turbine's safety.

[0077] (2) Energy storage grid support: Upon detecting a frequency drop, immediately operate at maximum power. Output: When a 220kV busbar in City A detects a frequency drop to 49.2Hz (below the rated value of 49.5Hz), the energy storage converter in the substation automatically switches to grid-connected mode within 5 milliseconds to discharge at maximum power. =50 MW Power is injected into the grid, pushing the frequency back to 49.6Hz.

[0078] (3) Distribution Network Islanding Supply Guarantee: The main grid fault triggers the preset microgrid, which is powered by distributed power sources and energy storage. When a 110kV line of the main grid trips due to a tower collapse fault, the distribution network automation terminal in the power supply area of ​​the line automatically detects the disconnection from the main grid. According to the preset microgrid operation strategy, some non-critical loads are cut off, and the 2MW distributed photovoltaic + 5MW / 10MWh energy storage system in the area forms an islanded microgrid to continuously supply power to the hospital (important user) and communication base station (extremely important user) in the area.

[0079] (4) Automatic load shedding: When power is insufficient When the load capacity reaches 120MW and the reserve capacity is exhausted, the station-level terminal executes the automatic load shedding command according to the contract priority. In this embodiment, the interruptible loads stipulated in the contract (response time < 0.5 seconds) are shedding first, totaling 80MW. The remaining 40MW shortfall is handled by the load naturally responding through voltage reduction, thus avoiding large-scale load shedding.

[0080] Step 4.5: Phase End Determination: When the typhoon wind speed remains below the threshold and the main faults are cleared, the process transitions to the recovery phase. The city cluster-level coordination center continuously monitors the typhoon wind speed. When the typhoon center moves out of the city cluster area, and the measured wind speed at all stations remains below 20 m / s for 30 minutes (below the equipment damage threshold), and the main fault points have been isolated, the city cluster-level coordination center determines that the response phase has ended and issues a phase switching instruction of "current phase = recovery phase" to each city-level dispatch center. Fault records and system status changes during the response phase (including a list of outages for each device, remaining energy storage capacity, available generation capacity, etc.) are saved and used as initial conditions for the recovery phase.

[0081] Step S5 (Black Start Optimization and Orderly Power Restoration during Recovery Phase): When the phase switching instruction received by the city-level dispatch center is for the recovery phase, the city-level dispatch center, based on the system state after the response phase, statistically analyzes the equipment outage status, remaining energy storage capacity, and available power generation capacity to construct a black start recovery optimization model with the goal of minimizing the maximum recovery time of the entire system. The city-level dispatch center prioritizes calling energy storage systems that have been forcibly fully charged in the prevention and control instructions and whose energy was not depleted during the response phase as black start power sources, and generates black start recovery instructions. The city-level dispatch center issues the black start recovery instructions to the station-level execution terminal for execution, establishes an isolated microgrid, and restores distributed power sources and distribution network lines in sequence. The city-level dispatch center also solves the cross-regional power support scheme through ADMM coordination between the city cluster-level coordination center and healthy cities. The city-level dispatch center puts in cold loads in batches according to the impact coefficient and importance level, and after each batch is put in, it verifies the voltage and frequency stability through real-time data fed back by the station-level execution terminal. Specifically, in a concrete implementation, this step can be: Sub-step 5.1 Disaster Loss Statistics and Initial State Modeling: Based on the real-time topology coloring function of SCADA / EMS, each city-level dispatch center statistically analyzes the disaster damage caused by the typhoon. Taking city A as an example, the statistical results are as follows: 220kV lines failed, 5 110kV lines failed, and 8 35kV lines failed; available generating capacity decreased from 2800MW to 1800MW (loss of 1000MW, including 300MW wind power loss, 2 thermal power units tripping for a total of 600MW, and other losses of 100MW); remaining energy storage capacity: energy storage station A has 45MWh remaining (200MWh fully charged, 155MWh discharged during the response phase), energy storage station B has 32MWh remaining, and energy storage station C has 28MWh remaining.

[0082] Sub-step 5.2 Black start recovery optimization: City-level dispatch center A constructs a black start recovery optimization model with the objective of minimizing the maximum recovery time of the entire system.

[0083] in, This refers to a collection of energy storage devices with black start capability (in this embodiment, it includes 3 energy storage power stations in city A, 2 energy storage power stations in city B, and 1 energy storage power station in city C). The recovery time for the b-th energy storage unit (the time from startup to successful establishment of a stable microgrid for that energy storage unit). The constraints include: (Condition 1) Minimum energy requirement for black start of energy storage: As an example, =10MWh, meaning the remaining energy storage capacity must be at least 10MWh. MWh Only then can it serve as a black-start power source. In this embodiment, the remaining power of the three energy storage power stations in City A is all greater than 10MWh, which meets the requirements.

[0084] (Condition 2) Reactive power limitation of line charging: ,in For the line l The charging reactive power, δ l A 0-1 variable indicates whether the route is included in the recovery path. This is the reactive power that the black-start power supply can provide. This constraint prevents excessive charging reactive power during recovery from causing voltage exceedances.

[0085] (Condition 3) Voltage dip limit for cooling load input: ,in To determine the bus voltage drop after a batch of cold loads is applied, in practice... =5% is the maximum allowable drop.

[0086] Sub-step 5.3 Layered recovery execution and cross-region support: In the specific implementation, the city-level dispatch center of city A generates a black start recovery command and issues it to the station-level execution terminal: (1) Priority Black Start of Energy Storage: Prioritize the use of energy storage systems that have been forcibly fully charged in step S3 and have not been depleted in step S4; as an example, energy storage power station A in city A has 45MWh remaining, which meets the black start requirement. The converter of energy storage power station A completes self-synchronization voltage building within 300 milliseconds after the command is issued, and establishes a microgrid voltage and frequency of 350V / 50Hz.

[0087] (2) Gradually restore distributed power sources: After the microgrid voltage stabilizes, the distributed photovoltaic power stations in the microgrid are put into operation in sequence (1MW is put into operation first, and the remaining part is put into operation after verifying the stability). The photovoltaic power stations automatically connect to the grid and generate electricity by detecting the existence of the grid.

[0088] (3) Distribution network line restoration step by step: The “charging-confirmation-expansion” strategy is adopted. After each level of line is restored, the bus voltage is checked to be within the range of 0.95~1.05pu before the next level of line is restored.

[0089] (4) Cross-regional power support: The ADMM coordination is restarted through the city cluster-level coordination center, and the cross-regional power support scheme is solved with the goal of minimizing the recovery time. In this embodiment, city B has approximately 150MW of remaining available power, and 80MW of power is transmitted to city A through the AB link to accelerate the recovery process of city A.

[0090] Sub-step 5.4: Phased activation of cooling load: During the load recovery phase, the station-level terminals are put into cold load in batches according to the impact coefficient and importance level. The impact coefficient refers to the current impact multiple at the moment the load is put into operation (the impact coefficient of motor loads can reach 5-8 times). As a specific implementation method, the recovery strategy is as follows: The first batch: for extremely important loads (hospitals, communication base stations, important institutions and departments), the impact factor can be set to be controlled within 2 times, and the expected voltage drop should be checked to be ≤3% before being put into operation; The second batch: For critical loads (large commercial and transportation facilities), the impact factor can be set to be controlled within 3 times; The third batch: ordinary loads (residential, general industrial), the impact factor can be set to be controlled within 4 times.

[0091] After each batch of power-on is activated, the station-level terminal collects bus voltage and frequency data in real time and transmits it back to the city-level dispatch center. After verification and confirmation that the voltage drop is ≤5% and the frequency deviation is ≤0.2Hz, the next batch of power-on instructions is executed. In this embodiment, it took 4.5 hours for City A to go from black start to restoring 80% load, which is about 40% shorter than the traditional restoration method (which usually takes 6-8 hours).

[0092] Step S6 (Operational Assurance and Smooth Switching): Throughout the typhoon event, the city cluster-level coordination center continuously monitors its communication status with each city-level dispatch center. When the city cluster-level coordination center is detected to be paralyzed, each city-level dispatch center automatically switches to its local pre-set conservative strategy to maintain minimum safe operation. After communication is restored, the state is smoothly synchronized and ADMM coordination is restarted. The city-level dispatch center also displays a probabilistic auxiliary decision panel to the dispatcher, which includes the optimal solution, success probability, robust radius, and list of weak nodes. It responds to the dispatcher's online adjustment instructions for optimization parameters and triggers re-optimization. After the typhoon passes, the city cluster-level coordination center performs post-event settlement of cross-regional support based on the ADMM multiplier trajectory and feeds back the model parameters and effects of this event to the historical database. Specifically, this step may further include the following sub-steps: Sub-step 6.1 Extreme Failure Autonomy: The city cluster-level coordination center continuously monitors its communication status with each city-level dispatch center. In this embodiment, based on the previous example, the dispatch center of city C experienced a temporary loss of contact during the response phase, triggering the following autonomy mechanism: When the city cluster-level coordination center detects a 30-second timeout in the heartbeat signal with the C city dispatch center, it immediately issues a takeover instruction to the B city dispatch center and simultaneously broadcasts a "Superior dispatch connection lost, switch to local preset conservative strategy" instruction to all station-level execution terminals under the jurisdiction of C city. Upon receiving this instruction, the station-level execution terminals in the C region automatically switch to the preset local conservative strategy: the island protection setting in their respective regions is reduced by 20%, energy storage switches to local VF control mode to maintain island voltage, and interruptible loads are disconnected locally with the lowest priority. When the C city communication link is restored after 45 minutes, the city cluster-level coordination center detects the heartbeat restoration and issues a "restore coordination mode" instruction. The C city-level dispatch center retrieves historical status data from each station-level terminal, smoothly synchronizes, and rejoins the ADMM coordination.

[0093] Sub-step 6.2 Human-Machine Collaborative Assisted Decision Making: Throughout the decision-making process, each city-level dispatch center displays a probabilistic assisted decision-making panel to the dispatchers. As an example, taking the display screen of the dispatcher's workstation in City A, the panel content may include: Summary of the optimal solution: "Solution #3 (ε-constrained scheme) is recommended, with an expected load shedding of 156 MWh." Success probability: "This solution can achieve a loss of ≤200MWh in 90% of typhoon scenarios"; Robust radius: =0.25, meaning that the loss can still be guaranteed to be ≤200MWh within ±25% of the actual wind speed deviation; List of vulnerable points: "High-risk equipment: L102 line (failure probability 0.72), G203 unit (N-1 pass rate 89%), B5 substation (total outage risk 0.15)"; Based on this, dispatchers can intervene manually by considering objective factors (such as a major event being held in a certain area requiring priority power supply): adjusting the ε2 value (safety constraint) or the IGDT loss baseline in the ε-constraint via the slider on the panel. Clicking the "Re-optimize" button triggers online re-optimization. The city-level dispatch center responds to the dispatcher's instruction, completing the re-optimization within 5 seconds and displaying the new solution, achieving a closed loop of human-machine collaboration.

[0094] Sub-step 6.3 Post-event settlement and experience accumulation: Settle cross-regional support based on ADMM multiplier trajectory (shadow price), and feed back model parameters and effects during the typhoon to the historical database to optimize the scenario probability distribution and robust parameters for the next event; As a specific implementation method, after the typhoon passes and the recovery phase ends, the city cluster-level coordination center will perform post-event settlement operations: (1) Cross-regional support settlement: based on the multiplier trajectory recorded during ADMM iteration. This multiplier has economic significance, reflecting the "shadow price" (i.e., marginal cost) of inter-regional support. In this embodiment, city B provided city A with a cumulative inter-regional support of 80MW×6h=480MWh. Based on the average shadow price of 0.35 yuan / kWh, city A should pay city B a support compensation fee of 168,000 yuan.

[0095] (2) Experience feedback: The entire process data of this typhoon event—including the deviation between the forecast scenario set and the actual typhoon path, the difference between the actual loss and the optimization target value at each stage, the convergence speed of ADMM, the actual execution effect of each control command, etc.—is packaged and stored in the historical database.

[0096] (3) Parameter optimization: Optimize the scenario probability distribution of the next event using feedback data (for example, if it is found that the actual path probability distribution of this typhoon is biased towards the southwest direction of the nominal distribution, then adjust the parameters in the next event). The weights) and robustness parameters (e.g., if the actual loss exceeds...). Then adjust κ or (The value of ).

[0097] Thus, this method embodiment completes a full closed-loop decision-making process for typhoon events, namely, from pre-disaster prevention, in-disaster response, post-disaster recovery to post-disaster feedback, providing a continuously optimized data foundation for decision-making on the next typhoon event.

[0098] The technical advantages of this application's embodiments are as follows: Through hierarchical collaboration between the city cluster-level coordination center, the city-level dispatch center, and the station-level execution terminal, a closed-loop decision-making process for the entire typhoon process is achieved: In the prevention phase, robust optimization using IGDT generates a pre-dispatch scheme that can withstand forecast errors, reserving adjustment margins for subsequent phases; in the response phase, rolling emergency control using MPC and dynamic management of failed areas minimizes real-time losses; and in the recovery phase, pre-set energy storage black-start resources are used to accelerate power restoration. All three phases of instructions are executed and fed back by the station-level terminal, forming a "decision-execution-feedback" closed loop. This method significantly reduces the risk of large-scale power outages in city clusters, shortens recovery time by 30%-50%, improves robustness to typhoon uncertainties, and the ADMM distributed architecture protects the data privacy of each city, is compatible with the existing dispatch system, and has high engineering practicality.

[0099] Secondly, such as Figure 3 As shown, the present invention also provides a source-grid-load-storage collaborative active defense and control system, which includes a city cluster-level coordination center, a city-level dispatch center, and a station-level execution terminal; the city cluster-level coordination center, the city-level dispatch center, and the station-level execution terminal cooperate with each other to execute the steps of the method described in the first aspect above.

[0100] For example, the city cluster-level coordination center in this embodiment can be a server (one or more) deployed at the superior dispatching agency; the city-level dispatching center in this application embodiment can be a server deployed at the dispatching agencies of each city (one set for each city); the station-level execution terminal in this application embodiment can be an edge intelligent terminal (such as a PLC, RTU, or protection device) deployed at substations, power plants, energy storage stations, and other sites.

[0101] Thirdly, embodiments of this application also protect a computer device, such as Figure 4 As shown, the computer device includes: at least one processor 10, a memory 11, and a computer program 12 stored in the memory 11 and executable on the at least one processor 10. When the processor 10 executes the computer program 12, it implements the steps in the above-described embodiment of the source-grid-load-storage coordinated active defense and control method of the first aspect of this application.

[0102] Corresponding to the embodiments described in the second aspect above, the computer device may be a server deployed on the upper-level dispatching agency (representing the aforementioned city cluster-level coordination center). The computer equipment may also be servers deployed in the dispatch agencies of each city (one set for each city, representing the aforementioned city-level dispatch center). The computer equipment can also be an edge intelligent terminal (such as PLC, RTU, protection device, representing the aforementioned station-level execution terminal) deployed at substations, power plants, energy storage stations, etc.

[0103] Figure 4 The computer device shown may include, but is not limited to, processor 10 and memory 11. Those skilled in the art will understand that... Figure 4 The computer device 1 is merely an example and does not constitute a limitation on the computer device 1. It may include more or fewer components than shown in the figure, or combine certain components, or different components, such as input / output devices, network access devices, etc.

[0104] This application also provides a computer program product, which, when executed by a processor, implements the steps in the various method embodiments described above.

[0105] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0106] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A source-grid-load-storage coordinated active defense and control method, characterized in that, The method is applied to a source-grid-load-storage coordinated active defense and control system, which includes a city cluster-level coordination center, a city-level dispatch center, and station-level execution terminals; the method includes: Step S1: The city cluster-level coordination center acquires the operation status data, equipment vulnerability parameters, and typhoon forecast scenario sets of the power grids of each city within the city cluster; the city cluster-level coordination center uses the operation status data, equipment vulnerability parameters, and typhoon forecast scenario sets belonging to different risk types in the power grids of each city as heterogeneous risk elements, uniformly maps them to the expected load loss risk value RaV of each city, and constructs the typhoon spatiotemporal stochastic evolution scenario set and its probability distribution uncertainty set; the city cluster-level coordination center also initializes the alternating direction multiplier method (ADMM) communication parameters between itself and the city-level dispatch center. Step S2: The city cluster-level coordination center automatically identifies the current stage type based on the time tag and geofence information in the typhoon spatiotemporal random evolution scenario set. The stage type includes prevention stage, response stage, and recovery stage. The city cluster-level coordination center uniformly issues stage switching instructions to each city-level dispatch center. Step S3: When the phase switching instruction received by the city-level dispatch center is the prevention phase, the city-level dispatch center, based on the expected load loss risk value (RaV), constructs a multi-objective Pareto optimization model with four-dimensional objectives: minimizing the total expected load loss of the system, minimizing the maximum risk range of the system, minimizing the total economic cost, and minimizing the recovery time; the city-level dispatch center uses the information gap decision theory (IGDT) to set the key typhoon parameters in the typhoon forecast scenario set as uncertain variables, and solves for a robust prevention dispatch scheme that can withstand the maximum forecast error; the city The city-level dispatch center uses the ADMM distributed iterative solution initialized with the ADMM communication parameters to solve the local optimization subproblems in parallel, gradually converging to the global Pareto optimal solution, while only exchanging the expected power values ​​of the tie lines with the city-level coordination center. The city-level dispatch center finally generates preventive control instructions, which include topology reconfiguration, forced full charging of energy storage, unit hot standby switching, and load peak avoidance. The city-level dispatch center issues the preventive control instructions to the station-level execution terminals for execution. The preventive control instructions are used to reserve adjustment margins for the subsequent response phase. Step S4: When the phase switching instruction received by the city-level dispatch center is in the response phase, the city-level dispatch center, based on the energy storage forced full charge state and unit hot standby switching state configured in the prevention and control instruction, collects wide-area measurement signals and fault signals in real time at fixed time periods to detect the communication and functional status of each area; for the detected failure areas, the city cluster-level coordination center automatically dynamically divides the failure areas according to the electrical adjacency matrix and entrusts them to the city-level dispatch centers of adjacent healthy cities; the city-level dispatch center adopts the model predictive control (MPC) method to minimize the load loss, frequency deviation, and voltage deviation in the current and predicted time domains, and solves the emergency control instructions in a rolling manner. The emergency control instructions include new energy power reduction flight, energy storage network support, distribution network island supply guarantee, and automatic load shedding; the city-level dispatch center issues the emergency control instructions to the station-level execution terminal for execution, and only executes the emergency control instructions of the current time step, and repeats the rolling optimization in the next cycle; the city-level dispatch center uses the fault records and system state changes in the response phase as the initial conditions for the recovery phase; Step S5: When the phase switching instruction received by the city-level dispatch center is the recovery phase, the city-level dispatch center, based on the system state after the response phase, statistically analyzes the equipment outage status, remaining energy storage capacity, and available power generation capacity, and constructs a black-start recovery optimization model with the goal of minimizing the maximum recovery time of the entire system; the city-level dispatch center prioritizes calling the energy storage system that has been forcibly fully charged in the prevention and control instruction and whose energy was not depleted in the response phase as the black-start power source, and generates a black-start recovery instruction; the city-level dispatch center issues the black-start recovery instruction to the station-level execution terminal for execution, establishes an islanded microgrid, and restores distributed power sources and distribution network lines in sequence; the city-level dispatch center also solves the cross-regional power support scheme through ADMM coordination between the city cluster-level coordination center and healthy cities; the city-level dispatch center puts in cold loads in batches according to the impact coefficient and importance level, and after each batch is put in, it verifies the voltage and frequency stability through real-time data feedback from the station-level execution terminal; Step S6: Throughout the typhoon event, the city cluster-level coordination center continuously monitors its communication status with each city-level dispatch center. When the city cluster-level coordination center is detected to be paralyzed, each city-level dispatch center automatically switches to its local pre-set conservative strategy to maintain minimum safe operation. After communication is restored, the state is smoothly synchronized and ADMM coordination is restarted. The city-level dispatch center also displays a probabilistic auxiliary decision panel to the dispatcher, which includes the optimal solution, success probability, robust radius, and list of weak nodes. It responds to the dispatcher's online adjustment instructions for optimization parameters and triggers re-optimization. After the typhoon passes, the city cluster-level coordination center performs post-event settlement of cross-regional support based on the ADMM multiplier trajectory and feeds back the model parameters and effects of this event to the historical database.

2. The method as described in claim 1, characterized in that, In step S1, the set of spatiotemporal stochastic evolution scenarios of typhoons and its uncertain probability distribution set are constructed, including: The city cluster-level coordination center constructs a set of spatiotemporal random evolution scenarios for typhoons. Each scenario in the set includes temporal wind speed, spatial distribution of rainfall, and probability weights. The city cluster-level coordination center uses φ-divergence to construct a probability distribution uncertainty set: Where κ is the divergence threshold, which represents the allowable deviation between the actual probability distribution and the nominal probability distribution.

3. The method as described in claim 1, characterized in that, In step S3, the multi-objective Pareto optimization model uses the enhanced ε-constraint method to transform the three objectives other than the main objective of minimizing the total expected load loss of the system into constraints. By changing the ε value, a complete Pareto front solution set is generated so that the scheduler can select the corresponding scheme according to actual preferences.

4. The method as described in claim 1, characterized in that, In step S4, the model predictive control (MPC) method includes: At each sampling time k, solve the following rolling optimization problem: in, Let be the load shedding amount at time t. For frequency deviation, For critical bus voltage deviation; These are the weighting coefficients. The unit penalty coefficient for frequency deviation, This is the unit penalty factor for voltage deviation; The constraints of the rolling optimization problem include: System frequency response swing equation constraints: Where H is the system inertia constant and D is the load damping coefficient; Let be the change in mechanical power at time t. Let be the change in electromagnetic power at time t; Constraints on the dynamic equations of the energy storage state of charge (SOC): ,in, For charge and discharge efficiency; Let be the remaining charge of energy storage b at time t.

5. The method as described in claim 1, characterized in that, In step S5, the objective function of the black start recovery optimization model is: in, A collection of energy storage devices with black start capability. Let be the recovery time for the b-th energy storage unit.

6. The method as described in claim 1, characterized in that, In step S3, the convergence criterion for the ADMM distributed iterative solution is: Calculate the original residuals and dual residual ; in the original residual And dual residuals In the case of convergence, it is determined to be convergent. in, For tie-line power variables, It is a copy of its global variable, and k is the iteration round.

7. The method as described in claim 1, characterized in that, The specific method of dynamic partitioning in step S4 is as follows: The city cluster-level coordination center continuously monitors the heartbeat signals of each city-level dispatch center. When the heartbeat signal times out or the measurement freezes, indicating a loss of communication or function in the affected area, the city cluster-level coordination center determines the cause based on the electrical adjacency matrix. The lost-connection area is automatically assigned to the city-level dispatch center of a neighboring healthy city. The city-level dispatch center of the healthy city incorporates a simplified equivalent model of the lost-connection area into its local optimization subproblem and uses residual measurements as pseudo-boundary conditions for unified optimization.

8. A source-grid-load-storage coordinated active defense and control system, characterized in that, The source-grid-load-storage coordinated active defense and control system includes a city cluster-level coordination center, a city-level dispatch center, and a station-level execution terminal; the city cluster-level coordination center, the city-level dispatch center, and the station-level execution terminal cooperate with each other to execute the steps of the method as described in any one of claims 1 to 7.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1 to 7.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a computer device, implements the steps of the method as described in any one of claims 1 to 7.

Images