Double-layer coordinated power system recovery method and system based on multiple flexible resources

Abstract

The invention discloses a double-layer coordinated power system recovery method and system based on multiple flexible resources. The method comprises the following steps: acquiring basic parameters of a target power system and setting a line recovery rate; solving the upper-layer topology reconstruction model to obtain a generator starting sequence and a skeleton network reconstruction scheme containing a bus connection state; solving the lower-layer flexible resource scheduling model to obtain a load recovery amount and flexible resource scheduling power controlled by a corresponding bus connection state; performing joint solution on the double-layer model by combining a global active power balance equation to obtain a preliminary recovery scheme; and performing physical security verification on the preliminary scheme, and if the verification is not passed, constructing an integer cutting constraint and feeding back the integer cutting constraint to the joint solution for iteration until a final recovery scheme is output. According to the method, strong coupling cooperative scheduling of grid reconstruction and multi-type flexible resources is realized, and the feasibility of global optimization and real power grid landing under extreme disasters is ensured through physical verification and a mutual exclusion constraint iteration mechanism.

Description

Technical Field

[0001] This invention belongs to the field of power system orderly power supply restoration technology, and particularly relates to a two-layer coordinated power system restoration method and system based on multiple flexible resources. Background Technology

[0002] With the rapid development of control technology and the widespread application of various automated equipment in power systems, the ability of power systems to resist interference and faults has been significantly improved. However, due to the large-scale integration of intermittent and fluctuating renewable energy sources such as wind and solar power into the power system, the power system faces greater complexity and uncertainty, posing challenges to its operational safety. Therefore, even in the event of a controllable islanding fault, the power system still faces the risk of a complete blackout or power failure. Systematic research on power system recovery strategies that consider multiple flexible resources after a partial or complete power outage has significant theoretical and practical implications, as it can ensure power supply security, reduce unrestored loads, and minimize economic losses.

[0003] Existing regional power restoration technologies primarily focus on phased independent restoration, conventional network topology reconfiguration, and the utilization of local flexible resources. In practice, traditional solutions often divide power system restoration into three stages: black-start partitioning, network reconfiguration, and load restoration, addressing each stage individually or in a partially coordinated manner. Regarding network reconfiguration, complex network characteristic parameters such as degree and betweenness coefficients, and modular indices are typically used to partition black-start regions. Partitioning strategies are designed based on ordered binary decision graphs and propagation characteristics, and generator startup sequences and network reconfiguration are modeled using mixed-integer linear programming. Simultaneously, to address uncertainties and the utilization of flexible resources, existing solutions often employ heuristic algorithms or sampling methods to handle uncertainties in renewable energy and electric vehicle systems, and introduce single or partial flexible resources (such as considering only wind power or stationary energy storage systems) into the model to provide limited black-start support.

[0004] While the aforementioned existing technologies have advanced the power system recovery process to some extent, they still have significant technical limitations when facing extreme disaster scenarios. First, most existing studies have severed the physical connection between upper-level network topology reconstruction and lower-level load and resource scheduling, resulting in insufficient stage coordination and global optimization capabilities. The lack of strong cross-stage coupling coordination mechanisms leads to recovery schemes that are often locally suboptimal, making it difficult to achieve the global optimum for total system-level load recovery. Second, the collaborative mining of flexible resources is incomplete. Existing solutions fail to incorporate various flexible resources such as renewable energy, electric vehicles, and energy storage systems into a unified scheduling framework, and their power support capabilities during black start-up and their uncertainty hedging capabilities under multiple scenarios are not fully explored. Furthermore, consideration of physical topology and equipment recovery characteristics is lacking. Not only are the real recovery response characteristics of various loads, such as constant, ramp, impact, and flexible loads, not considered in detail, but there is also a lack of deep integration planning of network connectivity indicators at the grid level. In particular, the risk of line faults under extreme disaster environments has been neglected for a long time, resulting in recovery paths derived solely based on mathematical optimization often lacking feasibility for implementation in real physical power grids.

[0005] Therefore, in response to the technical pain points of existing power system restoration technologies, such as fragmented phase coordination, one-sided utilization of flexible resources, and lack of physical security verification, there is an urgent need for a global optimization scheme that can coordinate network topology reconstruction and joint scheduling of multiple types of flexible resources. Summary of the Invention

[0006] To address the technical problems in existing power system recovery strategies, such as fragmented phase coordination, one-sided utilization of flexible resources, and lack of physical security verification leading to difficulties in implementation, the present invention aims to provide a two-layer coordinated power system recovery method and system based on multiple flexible resources. This method aims to achieve global optimization of power grid topology reconstruction and massive flexible resources under extreme disaster scenarios, effectively ensuring the safety of black start schemes and maximizing load recovery in real physical power grids.

[0007] To achieve the above objectives, the upper-layer model introduces two types of network topology indices: bus accessibility index and shortest reachability distance index. Combining generator starting characteristics (including cold / hot start time constraints and ramp rate constraints) with the recovery characteristics of various loads, mixed-integer linear programming (MILP) is used to optimize the non-black-start generator starting sequence and skeleton network reconstruction. Simultaneously, line recovery rate is incorporated to quantify the risk of recovery failure caused by extreme weather and inherent equipment faults. In the lower-layer model, multiple scenario sets are generated based on actual meteorological and operational data. The black-start capabilities and power regulation characteristics of renewable resources, electric vehicle systems, and energy storage systems are collaboratively utilized. Scenario analysis is used to handle the uncertainty of flexible resources, optimizing the load recovery scheme with the goal of maximizing weighted recoverable load. The characteristic of flexible loads being able to be secondary-cut off is used to prioritize the recovery of critical loads and buses. For the collaboration and solution mechanism of the two-layer model, strict cross-layer coupling constraints are constructed, ensuring that the scheduling actions of lower-layer resources are controlled by the bus connectivity status of the corresponding nodes in the upper layer; this is then combined with the global active power balance equation for joint solution. More importantly, this invention introduces physical security verification of AC power flow and frequency dynamic stability after joint solution, and constructs mutually exclusive integer cutting constraints based on the verification results to feed back to the original model for iterative optimization until infeasible solutions are completely eliminated and the final optimal recovery scheme is output.

[0008] The present invention adopts the following technical solution.

[0009] A two-tiered coordinated power system restoration method based on multiple flexible resources includes: Obtain the basic parameters of the target power system and set the line recovery rate; Based on the aforementioned basic parameters and line recovery rate, an upper-level topology reconfiguration model is constructed. With the goal of maximizing the sum of the generator's expected output capacity and the bus topology index, the generator start-up sequence and the skeleton network reconfiguration scheme including the bus connectivity status are determined. Based on the bus connection status, a lower-level flexible resource scheduling model is constructed to maximize the total recoverable load and determine the load recovery amount and flexible resource scheduling power that are limited by the corresponding bus connection status. By combining the global active power balance equation, the upper-level topology reconfiguration model and the lower-level flexible resource scheduling model are jointly solved to obtain a preliminary recovery scheme. The global active power balance equation includes power terms of generator output power, load recovery amount, and flexible resource scheduling power. The preliminary recovery plan is physically verified. If the verification fails, an integer slicing constraint is constructed to exclude the currently infeasible combination of the bus connectivity state and the flexible resource scheduling power. The integer slicing constraint is then fed back to the joint solution for iteration until the final recovery plan that passes the verification is obtained.

[0010] Preferably, the basic parameters include: data on buses and lines in the target power system, rated capacity, start-up sequence and ramp rate of generators, installation location, charging and discharging efficiency and rated capacity of renewable energy, electric vehicles and energy storage systems, as well as historical meteorological data and electric vehicle operation data; The setting of the line recovery rate specifically includes: dividing the lines in the target power system into different categories of line sets according to the geographical environment and the degree of disaster impact of the lines, and setting different line recovery rates for different categories of line sets; wherein, the line recovery rate of at least one category of line sets is lower than the line recovery rate of another category of line sets.

[0011] Preferably, in the objective function of the upper-level topology reconstruction model: The expected output capacity of the generator is determined by the joint probability of the effective output capacity of the generator to be started and the line recovery rate of each line on the recovery path from the recovered subnet to the generator to be started. The bus topology indices include the reachability index and the shortest reachability distance index; wherein, the reachability index is used to characterize the degree of disruption to the overall connectivity of the system caused by removing the corresponding bus, and the shortest reachability distance index is used to characterize the change in the minimum connectivity path length between network buses before and after removing the corresponding bus.

[0012] Preferably, the constraints of the upper-level topology reconfiguration model include generator operation constraints and network topology reconfiguration physical and logical constraints; The generator operation constraints are used to limit the output power boundaries of the generator at different recovery time steps based on the start-up timing and ramp rate in the basic parameters. The physical logic constraints of the network topology reconstruction include the constraint of not repeating outages after recovery and the constraint of restoring connectivity of associated lines; wherein, the constraint of not repeating outages after recovery is used to limit the buses or lines that have completed the recovery operation to remain in operation during subsequent recovery time steps; the constraint of restoring connectivity of associated lines is used to limit the lines to be restored to be connected to at least one of the buses that have completed the recovery.

[0013] Preferably, before constructing the lower-level flexible resource scheduling model, the method further includes: processing the historical meteorological data and the electric vehicle operation data using a clustering algorithm to generate renewable energy output scenarios and electric vehicle change scenarios respectively; and multiplying the occurrence probabilities of a single scenario based on the renewable energy output scenarios and the electric vehicle change scenarios to construct a probability vector containing multiple comprehensive scenarios and the occurrence probabilities of each comprehensive scenario. In the lower-level flexible resource scheduling model, maximizing the total recoverable load means maximizing the weighted expected total recoverable load under different scenarios. The weighted expected total recoverable load is calculated jointly by the load recovery amount on each bus, the load weight of each bus, and the probability of occurrence of each comprehensive scenario in the probability vector.

[0014] Preferably, the constraints of the lower-level flexible resource scheduling model include cross-layer coupling constraints and flexible resource operation constraints; The cross-layer coupling constraint is used to limit the load recovery amount, energy storage system scheduling power, and electric vehicle scheduling power on the corresponding bus to be greater than or equal to zero only when the bus connectivity status of the bus corresponding to the upper-layer topology reconfiguration model decision is restored; if the bus connectivity status is not restored, the load recovery amount and the scheduling power of each flexible resource on the corresponding bus are constrained to zero. The constraints on flexible resource operation include the charging and discharging power boundary constraints, energy balance constraints, and upper and lower limits of the state of charge constraints for energy storage systems and electric vehicles, as well as the output boundary constraints that match the output of renewable energy with the corresponding comprehensive scenario.

[0015] Preferably, the flexible resource scheduling power specifically includes: renewable energy output power, energy storage system charging and discharging power, and electric vehicle charging and discharging power; The global active power balance equation is configured such that, in each bus and each recovery time step of the target power system, the sum of the generator output power, the renewable energy output power, the energy storage system discharge power, and the electric vehicle discharge power of the corresponding bus is equal to the sum of the load recovery amount, the energy storage system charging power, and the electric vehicle charging power.

[0016] Preferably, the physical security verification includes AC power flow limit exceeding verification and frequency dynamic stability verification; The construction of integer cutting constraints specifically includes: recording the current bus connectivity state that causes the verification to fail; assigning variables with a value of "connected" to a first set and variables with a value of "unconnected" to a second set; generating the integer cutting constraints to force that in the subsequent joint solution, at least one bus connectivity state in the first set changes to "unconnected" or at least one bus connectivity state in the second set changes to "connected".

[0017] Secondly, a two-tiered coordinated power system restoration system based on multiple flexible resources includes: The parameter setting module is used to obtain the basic parameters of the target power system and set the line recovery rate; The upper-level reconfiguration module is used to construct an upper-level topology reconfiguration model based on the basic parameters and line recovery rate, with the goal of maximizing the sum of the expected output capacity of the generator and the bus topology index, and to determine the generator start-up sequence and the skeleton network reconfiguration scheme including the bus connectivity status. The lower-level scheduling module is used to construct a lower-level flexible resource scheduling model based on the bus connection status, with the goal of maximizing the total recoverable load, and to determine the load recovery amount and flexible resource scheduling power limited by the corresponding bus connection status. The joint solution module is used to combine the global active power balance equation to jointly solve the upper-level topology reconfiguration model and the lower-level flexible resource scheduling model to obtain a preliminary recovery scheme. The global active power balance equation includes power terms of generator output power, load recovery amount and flexible resource scheduling power. The verification iteration module is used to perform physical security verification on the preliminary recovery scheme. If the verification fails, an integer cutting constraint is constructed to exclude the currently infeasible combination of the bus connectivity state and the flexible resource scheduling power. The integer cutting constraint is then fed back to the joint solution for iteration until the final recovery scheme that passes the verification is obtained.

[0018] Thirdly, a terminal includes a processor and a storage medium; The storage medium is used to store instructions; The processor is configured to operate according to the instructions to perform the steps of any of the described two-tier coordinated power system restoration methods based on multiple flexible resources.

[0019] Fourthly, a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the two-tier coordinated power system restoration method based on multiple flexible resources as described in the first aspect.

[0020] The beneficial effects of this invention are as follows: 1) Refine the load recovery characteristic model: comprehensively consider the differentiated recovery laws of constant load, ramp load, impact load (including cold load picking characteristics) and flexible load. The load recovery process is made more realistic through piecewise linearization modeling. Among them, the characteristic of flexible load that can be cut off twice can free up power for the recovery of critical loads and busbars, and accelerate the overall recovery process.

[0021] 2) Collaborative optimization of the entire recovery process and network topology: Construct a two-layer model to collaboratively optimize the non-black start generator start sequence, recovery path and load recovery, and introduce "accessibility index" and "shortest reachability distance index" to determine the backbone network; at the same time, incorporate line recovery rate to quantify fault risk and improve the success rate and topology rationality of the recovery plan.

[0022] 3) Fully explore the value of multiple flexible resources: Comprehensively and synergistically utilize the black start capabilities and power regulation characteristics of renewable energy (RES), electric vehicle systems (EVS), and energy storage systems (ESS) to provide additional black start resources in the early stages of recovery, and to smooth out the imbalance between load and power generation, thus significantly accelerating the recovery process.

[0023] 4) Improve the practicality and adaptability of the model: By using scenario analysis (generating multiple types of RES outputs and EVS vehicle quantity scenarios based on actual data) to handle the uncertainty of flexible resources, the model performs better in different time periods and power systems of different scales, and surpasses existing models in indicators such as power generation capacity, recovery success rate, and recoverable load.

[0024] 5) It breaks through the limitations of traditional mixed-integer linear programming, which only focuses on purely mathematical optimal solutions. After joint solution, it forcibly introduces physical safety checks for AC power flow and frequency dynamic stability. By constructing integer slicing constraints to exclude infeasible solutions and trigger re-iteration, a set of physical fault tolerance and correction mechanisms is established to ensure that the final recovery solution output has the ability to be physically implemented in real complex power grid environments. Attached Figure Description

[0025] Figure 1 This is a flowchart illustrating an embodiment of a two-layer coordinated power system recovery strategy that considers multiple flexible resource supports provided by the present invention. Figure 2 This is a schematic diagram of a specific embodiment of a two-layer coordinated power system recovery strategy that takes into account multiple flexible resource supports provided by the present invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of this invention. The embodiments described in this application are merely some embodiments of this invention, and not all embodiments. Based on the spirit of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of this invention.

[0027] Example 1: like Figure 1 As shown, the specific steps of the two-layer coordinated power system restoration method based on multiple flexible resources are as follows: Step 1, System Parameters and Basic Data Acquisition: Obtain the basic parameters of the target power system.

[0028] The basic parameters include: data on buses and lines in the target power system, rated capacity, start-up sequence and ramp rate of generators, installation location, charging and discharging efficiency and rated capacity of renewable energy, electric vehicles and energy storage systems, as well as historical meteorological data and electric vehicle operation data.

[0029] Specifically, the process of collecting the above system parameters and basic data can be achieved through the following sub-steps: Step 1.1: Collect basic parameters such as rated capacity, start-up sequence, ramp rate, and start-up time of the target power system's buses, lines, generators (e.g., black start / non-black start), and loads (e.g., constant / ramp / impact / flexible loads) through the power system's energy management system or dispatch center database. These parameters serve as the foundational data for constructing the system's restored topology and physical constraints.

[0030] Step 1.2: Record the installation location, charging and discharging efficiency, rated capacity, black start capability, and other equipment parameters of renewable energy (RES, such as wind power and photovoltaic), electric vehicles (EVs), and energy storage systems (ESS) to provide boundary conditions for the subsequent flexible resource scheduling model.

[0031] Step 1.3: Collect historical meteorological data (wind speed, sunlight) and electric vehicle operation data (changes in the number of vehicles) to provide data support for the generation of subsequent uncertain scenarios.

[0032] Step 2, Scenario Generation and Constraint Parameter Setting: Generate a flexible resource operation scenario based on the basic parameters and set the line recovery rate.

[0033] This step provides uncertainty handling and boundary conditions for constructing the lower-level flexible resource scheduling model, which can be achieved through the following sub-steps 2.1 and 2.2: Step 2.1: Scene Generation Clustering algorithms are used to process the historical meteorological data and the electric vehicle operation data to generate renewable energy output scenarios and electric vehicle change scenarios, respectively. Based on the renewable energy output scenarios and the electric vehicle change scenarios, the occurrence probabilities of individual scenarios are multiplied together to construct a probability vector containing multiple comprehensive scenarios and the occurrence probabilities of each comprehensive scenario.

[0034] Specifically, the above scene generation process is as follows: Historical data is processed using the k-means clustering algorithm. The normalized 24-hour power output curve is used as the input feature vector to generate 12 types of solar energy output scenarios (sunny, cloudy, and rainy days in spring / summer / autumn / winter) and 3 types of wind energy output scenarios (high-speed, medium-speed, and low-speed) as the renewable energy output scenarios. The changes in the number of electric vehicles (EVs) are combined as the electric vehicle change scenarios to determine the probability of occurrence of each individual scenario. Subsequently, based on the assumption that the changes in wind, solar, and EVs are independent of each other, the occurrence probabilities of individual scenarios are multiplied together to construct a comprehensive scenario occurrence probability vector. Each element in this probability vector represents the occurrence probability of a specific combination of scenarios.

[0035] Step 2.2: Setting Constraint Parameters Define the boundary conditions and constraint parameters required for the system recovery model, specifically including: (1) Network topology parameters: Based on the geographical environment and the degree of disaster impact of the line, the lines in the target power system are divided into different categories of line sets, and different line recovery rates are set for different categories of line sets; among them, the line recovery rate of at least one type of line set is lower than the line recovery rate of another type of line set.

[0036] In practice, taking the simulation of equipment uncertainty under extreme disasters as an example, for the set of critical lines located at the disaster center or crossing complex terrain, their line recovery rate is set to 50% to assess the system's extreme recovery capability after the core topology is damaged; the line recovery rate of the remaining lines is set to 99.9% according to the normal failure rate.

[0037] (2) Time and operating parameters: Set the time step (default 5min) and the total recovery time. In mathematical modeling, the total recovery time is divided into multiple recovery time steps t of equal length. The length of each recovery time step is the time step, which serves as the unified decision step for physical processes such as network topology state change, flexible resource scheduling, and generator ramping. At the same time, clarify the operating constraint parameters such as generator cold / hot start time threshold.

[0038] (3) Load and flexible resource parameters: Give critical loads (hospitals, fire stations) higher load weights. For example, set the load weight of critical loads to 10 to 100 times that of ordinary loads to ensure that hospitals and other facilities are given priority in power supply during optimization. At the same time, set the operating safety threshold parameters of EVs and ESS (such as the upper and lower limits of state of charge SOC, charging and discharging power boundaries) and the maximum reduction ratio of flexible loads.

[0039] Step 3: Upper-level model construction and skeleton network optimization: Based on the basic parameters and line recovery rate, construct an upper-level topology reconfiguration model. With the goal of maximizing the sum of the expected output capacity of the generator and the bus topology index, determine the generator start-up sequence and the skeleton network reconfiguration scheme including the bus connectivity status.

[0040] Step 3 is described in detail below: Step 3.1: Calculate the network topology indices for each bus. The bus topology metrics include the reachability index and the shortest reachability distance index. The reachability index is used to characterize the degree of disruption to the overall connectivity of the system caused by removing the corresponding bus, and the shortest reachability distance index is used to characterize the change in the minimum connectivity path length between network buses before and after removing the corresponding bus.

[0041] Accessibility Index Shortest reachability index The specific calculation formula is as follows:

[0042]

[0043] in, It is the total number of buses in the target power system; This indicates whether bus m and bus n can be connected via a network; if connected, it is set to 1, otherwise it is set to 0. This indicates whether bus m and bus n can be connected through the network after bus i is removed. If the two buses can be connected, they are set to 1, otherwise they are set to 0. This item reflects the degree of damage to the overall connectivity of the system caused by removing bus i. This represents the shortest distance between bus m and bus n in the initial state; This represents the minimum distance between bus m and bus n after removing bus i; It is the set of bus pairs that can be reached after removing bus i.

[0044] The above accessibility index Shortest reachability index Together, they are used to characterize the increase in recovery costs and the degree of disruption to connectivity after the busbar is removed.

[0045] Step 3.2: Construct the upper-level objective function In the objective function of the upper-level topology reconfiguration model: the expected output capacity of the generator is determined by the joint probability of the line recovery rate of each line on the recovery path from the recovered subnet to the generator to be started, based on the effective output capacity of the generator to be started.

[0046] Specifically, the objective function aims to maximize the sum of the total expected output capacity of the generators (integrated line recovery rate) and the bus topology parameters:

[0047] in, This represents the maximized value of the upper-level objective function; This represents the specific recovery path from the currently restored subnet to the generator g to be started. Only the recovery rate of the lines along this path affects the expected revenue of the generator. This refers to the total installed capacity of the system, which is the sum of the capacities of all generators in the entire network. This is a weighting coefficient, which needs to be set according to the actual focus.

[0048] and These are the generating and starting capacities of generator g, respectively; This represents the total installed capacity, which is the sum of the capacities of all generators in the entire network; It is the collection of generators within the target power system; It is a set of lines to be restored; It is the line during the restoration period. The line recovery rate; t represents the current recovery time step; This is a binary variable representing the connectivity state of bus i at recovery time step t. If bus i has been reconnected at recovery time step t, then... ,otherwise .

[0049] Step 3.3: Construct upper-level constraints The constraints of the upper-level topology reconfiguration model include generator operation constraints and network topology reconfiguration physical and logical constraints. The generator operation constraints are used to limit the output power boundaries of the generator at different recovery time steps based on the start-up timing and ramp rate in the basic parameters. The physical logic constraints of the network topology reconstruction include the constraint of not repeating outages after recovery and the constraint of restoring connectivity of associated lines; wherein, the constraint of not repeating outages after recovery is used to limit the buses or lines that have completed the recovery operation to remain in operation during subsequent recovery time steps; the constraint of restoring connectivity of associated lines is used to limit the lines to be restored to be connected to at least one of the buses that have completed the recovery.

[0050] In practical implementation, the mathematical model for the above constraints is as follows: (1) Generator operating constraints:

[0051]

[0052] in, It is a collection of generator sets; , and These are the rated starting power, climbing rate, and maximum output power of generator g, respectively. and These are the starting power and output power of generator g during recovery time t, respectively. , and These are the start-up time, the start-up time, and the time when the power output of generator g reaches its maximum value, which together constitute the start-up sequence. This is the recovery time required for generator g to start.

[0053] (2) Physical and logical constraints of network topology reconstruction: In addition, the upper-level model must also satisfy the physical and logical constraints of network topology reconstruction, specifically including: ① No repeated outages after line / busbar restoration: Once a busbar or line restoration operation is completed, it should remain operational during subsequent restoration cycles, and its status indicator variables should satisfy the following:

[0054] in, and These represent bus connection status indicator variables for bus i at the current recovery time step t and the previous recovery time step t-1, respectively. The value is 1 when the bus is connected and 0 when it is not connected. For the set of busbars in the target power system, This is the total recovery time, or the total time step. This constraint ensures that once the state variable of the bus changes from 0 to 1, it will remain at 1 in subsequent time periods, thus maintaining its operational status.

[0055] ② Reconnection constraints for associated lines: The line to be restored must be connected to at least one of its restored busbars to ensure the connectivity of the restoration path.

[0056] During the model solution phase, for the aforementioned nonlinear constraints involving logical judgments and piecewise linear functions in the generator model, the nonlinear constraints are linearized using the Big M method; alternatively, the piecewise linear functions are directly processed using the built-in General Constraint function of commercial solvers (such as Gurobi / CPLEX), whose underlying logic is based on state indicator variables and the Big M method to achieve linearization transformation. Finally, mixed-integer linear programming (MILP) is used to solve the problem, obtaining the optimal generator start-up sequence and skeleton network reconstruction scheme.

[0057] Step 4, Lower-level model construction and flexible resource collaborative modeling: Based on the bus connection status, a lower-level flexible resource scheduling model is constructed to maximize the total recoverable load and determine the load recovery amount and flexible resource scheduling power limited by the corresponding bus connection status.

[0058] Step 4 is as follows: Step 4.1: Construct the lower-level objective function Based on the skeleton network and generator startup sequence output from the upper layer, the objective function of the lower-layer flexible resource scheduling model is constructed. Maximizing the total recoverable load is achieved by maximizing the weighted expected total recoverable load under different scenarios. The weighted expected total recoverable load is jointly calculated from the load recovery amount on each bus, the load weight of each bus, and the probability of occurrence of each comprehensive scenario in the probability vector.

[0059] Maximizing the total weighted recoverable load in multiple scenarios:

[0060] Where S is the number of integrated scenarios; It is the probability of the s-th scenario occurring; To optimize the variables, we define the load recovery amount on bus i in scenario s at recovery time step t; the subscript s indicates the corresponding value in the s-th scenario. This represents the load weight of bus i, used to reflect the importance of the load on bus i.

[0061] Step 4.2: Establish a load recovery characteristic model The recovery curves of constant, ramp, and impact loads are described by piecewise linearization, and the flexible load can be used to adapt the model to power regulation requirements.

[0062]

[0063] in, , and These are the recovery time of bus i, the time for the impact load to decrease to the rated value of bus i, and the time for the ramp load to increase to the nominal value of bus i, respectively.

[0064] To achieve cross-level coordination and variable transfer in the two-level model, the recovery time of bus i... It is influenced by the state indicator variable of the upper topology. Strictly constrained correlation variables, i.e. for The time step when the connection status of the bus corresponding to the upper-level decision is not restored, i.e., when it first jumps from 0 to 1. The corresponding load recovery amount on the busbar It is constrained to zero.

[0065] and and These are inherent parameters related to the load, which can be determined based on actual conditions or estimated empirically. In practice, and It can be determined based on the actual situation or estimated based on experience. and These are the time values ​​of bus i. and The load; The maximum / rated load demand on bus i; and These are the sets of all busbars and the sets of busbars connected to flexible loads, respectively.

[0066] Step 4.3: Constructing Flexible Resource Constraints The constraints on flexible resource operation include the charging and discharging power boundary constraints, energy balance constraints, and upper and lower limits of the state of charge constraints for energy storage systems and electric vehicles, as well as the output boundary constraints that match the output of renewable energy with the corresponding comprehensive scenario.

[0067] Specifically, this includes constraints on ESS charging and discharging power and energy balance, constraints on EV charging and discharging characteristics and vehicle quantity uncertainty, and constraints on RES output and scenario matching, to ensure that multiple resources work together.

[0068] (1) Energy Storage System (ESS) Operational Constraints:

[0069]

[0070]

[0071]

[0072] in, , and These are the charging power, discharging power, and remaining energy of the ESS connected to bus i at recovery time t; , , These are the maximum charging power, maximum discharging power, and rated capacity of the ESS connected to bus i, respectively. and This refers to the charging and discharging efficiency of ESS; This refers to the time step of the operation. As a binary variable, it ensures that the energy storage system charges and discharges at different times.

[0073] (2) Operating constraints of electric vehicles (EVs):

[0074]

[0075]

[0076]

[0077] in, These are the charging power, discharging power, and remaining energy of the EVS connected to bus i at recovery time t; , , These are the maximum charging power, maximum discharging power, and rated capacity of a single EVS connected to bus i. and This refers to the charging and discharging efficiency of EVs; The number of electric vehicles at the station at time t; The energy consumed / taken away by the electric vehicle that leaves the station during recovery time t.

[0078] (3) Constraints on the matching of renewable energy (RES) output with the scenario: To address uncertainty, the actual dispatchable output power of renewable energy is limited not only by its installed capacity but also by the current recovery time step and the predicted maximum output boundary under the corresponding comprehensive scenario, satisfying:

[0079] in, It is the actual output power of renewable energy connected to bus i under recovery time t and comprehensive scenario s; It is the predicted maximum available output of renewable energy on bus i under recovery time t based on the corresponding comprehensive scenario s generated in step 2. This predicted upper limit determines the output boundary under the corresponding scenario.

[0080] Step 4.4: Construct cross-layer coupling constraints for the two-layer model The cross-layer coupling constraint is used to limit the load recovery amount, energy storage system scheduling power, and electric vehicle scheduling power on the corresponding bus to be greater than or equal to zero only when the bus connectivity status of the bus corresponding to the upper-layer topology reconfiguration model decision is restored; if the bus connectivity status is not restored, the load recovery amount and the scheduling power of each flexible resource on the corresponding bus are constrained to zero. To achieve strict physical coordination between upper-level network reconfiguration and lower-level resource scheduling, and to ensure the consistency of variable logic during joint solution using a single MILP model (step 5), all load restoration behaviors on lower-level bus i and the power scheduling of flexible resources (ESS / EVs / RES) must be constrained by the bus connectivity status output by the upper level. Introduce the following cross-layer coupling constraints for the Big M method:

[0081]

[0082]

[0083]

[0084] Among the above constraints, This refers to the state indicator variable of the upper-level model decision bus i at recovery time t. This set of constraints indicates that only if the upper-level model decision bus i has completed physical recovery at time t (i.e., Only when the busbar has not recovered ( ) is the lower-level model allowed to allocate load recovery and flexible resource output; if the busbar has not recovered ( ) If ), then the corresponding output and load are forcibly constrained to 0.

[0085] Step 5: Optimize and solve the load restoration scheme. Combining the global active power balance equation, the upper-level topology reconfiguration model and the lower-level flexible resource scheduling model are jointly solved to obtain a preliminary restoration scheme. The global active power balance equation includes power terms for generator output power, the load restoration amount, and the flexible resource scheduling power.

[0086] Step 5, as follows: Step 5.1: Add power balance constraints The global active power balance equation is configured such that, in each bus and each recovery time step of the target power system, the sum of the generator output power, the renewable energy output power, the energy storage system discharge power, and the electric vehicle discharge power of the corresponding bus is equal to the sum of the load recovery amount, the energy storage system charging power, and the electric vehicle charging power.

[0087] The specific mathematical expression is:

[0088] in, The output power of the generator connected to bus i during recovery time t (0 if there is no generator on the bus). To restore the output power of renewable energy connected to bus i under the comprehensive scenario s, time t; These are the real-time charging and discharging power of ESS, respectively; Real-time charging and discharging power of EVS; This represents the power exchanged with the upstream power grid (positive for buying electricity, negative for selling electricity). This represents the load recovery amount of bus i under recovery time t and comprehensive scenario s. It ensures the dynamic balance between bus-side power generation (generator + RES + ESS / EVS discharge + grid power purchase) and load consumption (including ESS / EVS charging) in each time period and scenario.

[0089] Step 5.2: Use scenario analysis to handle uncertainties in flexible resources The upper-level topology reconstruction model from step 3 is combined with the lower-level flexible resource scheduling model from step 4. By introducing a bus connectivity status indicator variable, the cross-layer coupling constraints are transformed into a single mixed-integer linear programming (MILP) model, which is then solved using a commercial solver (such as Gurobi / CPLEX) to obtain the optimal load recovery amount and flexible resource scheduling power under each integrated scenario.

[0090] Step 5.3: Output a preliminary recovery plan that includes generator start-up sequence, skeleton network reconfiguration scheme, load recovery amount of each bus, and flexible resource scheduling strategy.

[0091] Step 6: Scheme Verification and Final Determination. Perform physical security verification on the preliminary recovery scheme; if the verification fails, construct integer slicing constraints to exclude currently infeasible combinations of the bus connectivity state and the flexible resource scheduling power; and feed these integer slicing constraints back into the joint solution for iteration until a final recovery scheme that passes the verification is obtained.

[0092] Step 6 is as follows: Step 6.1: Using the preliminary recovery scheme output in Step 5 as input, perform the physical security verification. The physical security verification specifically includes AC power flow limit verification and frequency dynamic stability verification. (1) AC power flow limit verification: Run the AC power flow calculation program to calculate the voltage amplitude of each bus and the reactive power flow of the line, and verify whether there are voltage limit exceedance or reactive power deficiency problems.

[0093] (2) Frequency dynamic stability verification: Based on the system frequency response model (SFR) or time domain simulation, the frequency drop depth at the moment of generator start-up and load input is deduced to verify frequency stability and ensure that the scheme meets the power system operation safety requirements; Step 6.2: If the preliminary recovery scheme does not meet any of the above verification constraints, the current solution is determined to be infeasible, and the integer split constraint is triggered.

[0094] The construction of integer slicing constraints specifically includes: recording the current bus connectivity state that causes the verification to fail; assigning variables with a value of "connected" to a first set and variables with a value of "disconnected" to a second set; "connected" is indicated by a state indicator variable of 1, and "disconnected" is indicated by a state indicator variable of 0. Integer slicing constraints are generated using integer slicing techniques to force that in the subsequent joint solution, at least one bus connectivity state in the first set changes to "disconnected," or at least one bus connectivity state in the second set changes to "connected."

[0095] This constraint aims to forcibly exclude the combination of currently infeasible bus connectivity state and flexible resource scheduling power, that is, to prohibit the decision variable vector from taking this set of values ​​again. The constraint is then added back to the MILP model in step 5, triggering the solver to re-optimize, obtain the suboptimal solution in the remaining solution space, and repeat the verification process in step 6.1 until a feasible solution is obtained.

[0096] Step 6.3: Output the final recovery scheme that passes the verification, and clarify the key operations for each time period, including: generator start-up sequence, skeleton network reconfiguration scheme including bus connectivity status, load recovery amount on each bus and scheduling power of various flexible resources (ESS / EVs / RES).

[0097] Example 2: like Figure 2 As shown, a two-tiered coordinated power system restoration system based on multiple flexible resources includes: The parameter setting module is used to obtain the basic parameters of the target power system and set the line recovery rate; The upper-level reconfiguration module is used to construct an upper-level topology reconfiguration model based on the basic parameters and line recovery rate, with the goal of maximizing the sum of the expected output capacity of the generator and the bus topology index, and to determine the generator start-up sequence and the skeleton network reconfiguration scheme including the bus connectivity status. The lower-level scheduling module is used to construct a lower-level flexible resource scheduling model based on the bus connection status, with the goal of maximizing the total recoverable load, and to determine the load recovery amount and flexible resource scheduling power limited by the corresponding bus connection status. The joint solution module is used to combine the global active power balance equation to jointly solve the upper-level topology reconfiguration model and the lower-level flexible resource scheduling model to obtain a preliminary recovery scheme. The global active power balance equation includes power terms of generator output power, load recovery amount and flexible resource scheduling power. The verification iteration module is used to perform physical security verification on the preliminary recovery scheme. If the verification fails, an integer cutting constraint is constructed to exclude the currently infeasible combination of the bus connectivity state and the flexible resource scheduling power. The integer cutting constraint is then fed back to the joint solution for iteration until the final recovery scheme that passes the verification is obtained.

[0098] Based on Example 1, in order to further strengthen the cross-layer physical coupling between upper-layer topology reconstruction and lower-layer flexible resource scheduling, and to break the limitation of static and fixed load recovery priority in the traditional power grid recovery model, this example proposes a dynamic weighting method for lower-layer loads based on upper-layer network topology indicators.

[0099] Specifically, in the lower-level flexible resource scheduling model, the load weights of each bus used to calculate the total expected weighted recoverable load are configured as dynamic weights that change dynamically with the upper-level topology. The calculation formula is as follows:

[0100] In the formula, The static base weight of the load on bus i is determined by the social / economic importance of the load, such as hospitals and government agencies, which are assigned higher base weights. The bus reachability index corresponding to bus i output by the upper-level topology reconstruction model at recovery time step t; This is the weight adjustment coefficient for the topology index.

[0101] In scenarios where extreme cold disasters, such as freezing temperatures, cause significant changes to the target power system topology, the importance of a bus depends not only on its load characteristics but also on its spatial topological position within the current reconfigured power grid. Using the formula above, when the upper-level topology reconfiguration decision identifies a bus as a critical hub for system connectivity (i.e., an accessibility index)... When the value is extremely high, the indicator will be directly passed across layers to the lower layer, automatically amplifying the dynamic weight of the bus.

[0102] Example 3: In actual power system recovery processes during extreme disasters, not only are there uncertainties in meteorological condition predictions, but the bi-level mixed integer linear programming (MILP) model also suffers from the curse of dimensionality when dealing with large-scale power grids. Furthermore, the topology commands output by artificial intelligence also carry the risk of error. To overcome these shortcomings, this embodiment, based on Embodiment 1, proposes a data- and mechanism-driven dual-drive rolling control method based on confidence-based soft constraints, predictive cutting, and dynamic weight adjustment. The specific steps are as follows: Step A, Parameter Initialization and Online Feature Extraction: Define the current actual physical time step as Obtain the current damaged power grid topology map and weather forecast data.

[0103] Real-time data is input into a pre-trained graph neural network (GNN). The network outputs topology action labels and corresponding confidence scores for each bus to be restored; in addition, by combining historical solution failure patterns, the network simultaneously outputs high-risk feature vectors of potentially infeasible topology combinations.

[0104] Step B, Target Reconstruction and Predictive Cut Generation Based on Confidence: (1) Predictive cut generation: Change the traditional post-verification feedback mode. Based on the high-risk feature vector of the potential infeasible topology combination output in step A, predictive integer cut constraints are generated in advance before MILP solution, and infeasible regions that are prone to voltage collapse or frequency instability are cut off in advance in the solution space.

[0105] (2) Dynamic objective function reconstruction: Based on the topological action labels and confidence scores, the static load weights of each bus in the lower-level model are dynamically corrected to obtain dynamic load weights; for example, for buses marked as high-risk by GNN, their load weights are reduced, and for buses that are bound to be connected with high confidence, their load weights are increased.

[0106] Step C, construct confidence-driven soft constraints and joint solution: Based on the aforementioned topology action labels and confidence scores, the bus state indicator variables are divided into strong constraint variables and weak constraint variables. A soft constraint model that allows for violation is constructed: In the upper-level objective function of step 3, a deviation penalty term is introduced. This penalty term is constructed based on the confidence score, and is applied when the solver decides on the actual bus connectivity state. When the label deviates from the GNN output, an algebraic penalty positively correlated with the confidence level is applied. This mechanism allows the solver to break the limitations of the AI ​​label by incurring a penalty cost when encountering severe physical active power balance conflicts, thus ensuring the absolute feasibility of the underlying physical solution.

[0107] A commercial solver is used to perform joint optimization under the modified objective function and predictive cutting constraints, and the decision instruction for the first time step is extracted and issued for execution.

[0108] Step D, Status Feedback and Window Sliding: After the physical device executes the command, it collects the latest operating parameters. This sets the actual physical time step. If the predicted window slides, return to step A for the next round of rolling closed-loop control.

[0109] A terminal, comprising a processor and a storage medium; The storage medium is used to store instructions; The processor is configured to operate according to the instructions to perform the steps of the two-layer coordinated power system restoration method based on multiple flexible resources as described in Embodiment 1.

[0110] Example 4: A computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements the steps of the two-tier coordinated power system restoration method based on multiple flexible resources as described in Embodiment 1.

[0111] This disclosure can be a system, method, and / or computer program product. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of this disclosure.

[0112] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example—but not limited to—electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination of the foregoing. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0113] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0114] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.

[0115] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A two-layer coordinated power system restoration method based on multiple flexible resources, characterized in that, include: Obtain the basic parameters of the target power system and set the line recovery rate; Based on the aforementioned basic parameters and line recovery rate, an upper-level topology reconfiguration model is constructed. With the goal of maximizing the sum of the generator's expected output capacity and the bus topology index, the generator start-up sequence and the skeleton network reconfiguration scheme including the bus connectivity status are determined. Based on the bus connection status, a lower-level flexible resource scheduling model is constructed to maximize the total recoverable load and determine the load recovery amount and flexible resource scheduling power that are limited by the corresponding bus connection status. By combining the global active power balance equation, the upper-level topology reconfiguration model and the lower-level flexible resource scheduling model are jointly solved to obtain a preliminary recovery scheme. The global active power balance equation includes power terms of generator output power, load recovery amount, and flexible resource scheduling power. The preliminary recovery plan was physically verified. If the verification fails, an integer slicing constraint is constructed to exclude combinations of the bus connectivity state and the flexible resource scheduling power that are currently infeasible; and the integer slicing constraint is fed back to the joint solution for iteration until a final recovery scheme that passes the verification is obtained.

2. The two-layer coordinated power system restoration method based on multiple flexible resources according to claim 1, characterized in that: The basic parameters include: data on buses and lines in the target power system, rated capacity, start-up sequence and ramp rate of generators, installation location, charging and discharging efficiency and rated capacity of renewable energy, electric vehicles and energy storage systems, as well as historical meteorological data and electric vehicle operation data; The setting of the line recovery rate specifically includes: dividing the lines in the target power system into different categories of line sets according to the geographical environment and the degree of disaster impact of the lines, and setting different line recovery rates for different categories of line sets; wherein, the line recovery rate of at least one category of line sets is lower than the line recovery rate of another category of line sets.

3. The two-layer coordinated power system restoration method based on multiple flexible resources according to claim 2, characterized in that: In the objective function of the upper-level topology reconstruction model: The expected output capacity of the generator is determined by the joint probability of the effective output capacity of the generator to be started and the line recovery rate of each line on the recovery path from the recovered subnet to the generator to be started. The bus topology indices include the reachability index and the shortest reachability distance index; wherein, the reachability index is used to characterize the degree of disruption to the overall connectivity of the system caused by removing the corresponding bus, and the shortest reachability distance index is used to characterize the change in the minimum connectivity path length between network buses before and after removing the corresponding bus.

4. The two-layer coordinated power system restoration method based on multiple flexible resources according to claim 3, characterized in that: The constraints of the upper-level topology reconfiguration model include generator operation constraints and network topology reconfiguration physical and logical constraints. The generator operation constraints are used to limit the output power boundaries of the generator at different recovery time steps based on the start-up timing and ramp rate in the basic parameters. The physical logic constraints of the network topology reconstruction include the constraint of not repeating outages after recovery and the constraint of restoring connectivity of associated lines; wherein, the constraint of not repeating outages after recovery is used to limit the buses or lines that have completed the recovery operation to remain in operation during subsequent recovery time steps; the constraint of restoring connectivity of associated lines is used to limit the lines to be restored to be connected to at least one of the buses that have completed the recovery.

5. The two-layer coordinated power system restoration method based on multiple flexible resources according to claim 2, characterized in that: Before constructing the lower-level flexible resource scheduling model, the method further includes: using a clustering algorithm to process the historical meteorological data and the electric vehicle operation data to generate renewable energy output scenarios and electric vehicle change scenarios respectively; based on the renewable energy output scenarios and the electric vehicle change scenarios, multiplying the occurrence probability of a single scenario to construct a probability vector containing multiple comprehensive scenarios and the occurrence probability of each comprehensive scenario. In the lower-level flexible resource scheduling model, maximizing the total recoverable load means maximizing the weighted expected total recoverable load under different scenarios. The weighted expected total recoverable load is calculated jointly by the load recovery amount on each bus, the load weight of each bus, and the probability of occurrence of each comprehensive scenario in the probability vector.

6. The two-layer coordinated power system restoration method based on multiple flexible resources according to claim 5, characterized in that: The constraints of the lower-level flexible resource scheduling model include cross-layer coupling constraints and flexible resource operation constraints. The cross-layer coupling constraint is used to limit the load recovery amount, energy storage system scheduling power, and electric vehicle scheduling power on the corresponding bus to be greater than or equal to zero only when the bus connectivity status of the bus corresponding to the upper-layer topology reconfiguration model decision is restored; if the bus connectivity status is not restored, the load recovery amount and the scheduling power of each flexible resource on the corresponding bus are constrained to zero. The constraints on flexible resource operation include the charging and discharging power boundary constraints, energy balance constraints, and upper and lower limits of the state of charge constraints for energy storage systems and electric vehicles, as well as the output boundary constraints that match the output of renewable energy with the corresponding comprehensive scenario.

7. The two-layer coordinated power system restoration method based on multiple flexible resources according to claim 1, characterized in that: The flexible resource dispatch power specifically includes: renewable energy output power, energy storage system charging and discharging power, and electric vehicle charging and discharging power; The global active power balance equation is configured such that, in each bus and each recovery time step of the target power system, the sum of the generator output power, the renewable energy output power, the energy storage system discharge power, and the electric vehicle discharge power of the corresponding bus is equal to the sum of the load recovery amount, the energy storage system charging power, and the electric vehicle charging power.

8. The two-layer coordinated power system restoration method based on multiple flexible resources according to claim 1, characterized in that: The physical security verification includes AC power flow limit exceeding verification and frequency dynamic stability verification; The construction of integer cutting constraints specifically includes: recording the current bus connectivity state that causes the verification to fail; assigning variables with a value of "connected" to a first set and variables with a value of "unconnected" to a second set; generating the integer cutting constraints to force that in the subsequent joint solution, at least one bus connectivity state in the first set changes to "unconnected" or at least one bus connectivity state in the second set changes to "connected".

9. A two-tiered coordinated power system restoration system based on multiple flexible resources, utilizing the method described in any one of claims 1-8, characterized in that, include: The parameter setting module is used to obtain the basic parameters of the target power system and set the line recovery rate; The upper-level reconfiguration module is used to construct an upper-level topology reconfiguration model based on the basic parameters and line recovery rate, with the goal of maximizing the sum of the expected output capacity of the generator and the bus topology index, and to determine the generator start-up sequence and the skeleton network reconfiguration scheme including the bus connectivity status. The lower-level scheduling module is used to construct a lower-level flexible resource scheduling model based on the bus connection status, with the goal of maximizing the total recoverable load, and to determine the load recovery amount and flexible resource scheduling power limited by the corresponding bus connection status. The joint solution module is used to combine the global active power balance equation to jointly solve the upper-level topology reconfiguration model and the lower-level flexible resource scheduling model to obtain a preliminary recovery scheme. The global active power balance equation includes power terms of generator output power, load recovery amount and flexible resource scheduling power. The verification iteration module is used to perform physical security verification on the preliminary recovery plan; If the verification fails, an integer slicing constraint is constructed to exclude the currently infeasible combination of the bus connectivity state and the flexible resource scheduling power. The integer slicing constraint is then fed back to the joint solution for iteration until a final recovery scheme that passes the verification is obtained.

10. A terminal, comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is configured to operate according to the instructions to perform the steps of the two-tier coordinated power system restoration method based on multiple flexible resources as described in any one of claims 1-8.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the steps of the two-tier coordinated power system restoration method based on multiple flexible resources as described in any one of claims 1-8.

Images