A method for balancing power of public power grid in n-m scenario

Abstract

This invention discloses a method for analyzing the power balance of a public power grid in an n-m scenario, comprising the following analysis steps: S1, constructing an event-driven model based on the event chain and the trigger-propagation-termination rule of time; S2, constructing a spatiotemporal probability balance analysis model for power supply in different scenarios; S3, generating the association pattern of events based on the event-driven model, and generating high-dimensional scenario samples through Vines-Copula; S4, solving the power balance equation for each sample in different time periods to obtain the power deficit / surplus and the power supply-demand difference at each time; S5, statistically analyzing the calculation results of all samples to obtain the probability distribution and risk indicators of power imbalance; S6, constructing a three-level defense strategy of prevention-emergency-recovery based on the risk indicators output by the model; This invention can realize the analysis and quantitative evaluation of the power balance state under multiple event disturbances, ensuring the safe and stable operation of the power system.

Description

Technical Field

[0001] This invention belongs to the field of power balance analysis technology for public power grids, and specifically relates to a method for power balance analysis of public power grids in a nm scenario. Background Technology

[0002] With the accelerated advancement of energy transition, a large number of distributed power sources (such as photovoltaic and wind power) and energy storage systems are connected to the public power grid, forming a complex operation pattern of integrated source, grid, load and storage. While this transformation improves energy utilization efficiency and promotes the consumption of clean energy, it also brings many challenges to the voltage security of the public power grid.

[0003] The nm scenario refers to the situation where m components (such as lines, transformers, power sources, etc.) fail and go out of operation simultaneously or successively in a power system. The nm scenario will disrupt the original power balance of the public power grid, causing the power distribution to be readjusted, and the coupling effect between events (concurrency, sequential generation, cascading) will exacerbate the uncertainty of the system. At the same time, the intermittency of distributed power sources, the randomness of loads, and the regulation capability of energy storage systems make the power balance relationship of the public power grid more complex.

[0004] Therefore, in order to solve the above problems, it is necessary to develop a power balance analysis method for public power grids in the nm scenario. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method for analyzing the power balance of a public power grid in a nm scenario, so as to realize the analysis and quantitative evaluation of the power balance state under multiple event disturbances and ensure the safe and stable operation of the power system.

[0006] The objective of this invention is achieved as follows: a method for analyzing the power balance of a public power grid in a nm scenario, comprising the following analysis steps:

[0007] S1. Based on the event chain as the core and the trigger-propagation-termination rule of time, an event-driven model is constructed, which includes event attribute definition, scenario classification and correlation strength quantification. The scenario classification includes concurrent event scenarios, sequential event scenarios and cascading event scenarios.

[0008] S2. Construct a spatiotemporal probability balance analysis model for power consumption based on different scenarios, specifically including event concurrency scenario model, event sequential scenario model, and event cascading scenario model;

[0009] S3, generated based on event-driven model The correlation patterns of events are analyzed, and high-dimensional scene samples are generated using vines-Copula. Each sample contains event sequences, power perturbations, and network parameters.

[0010] S4. For each sample, solve the power balance equation in different time periods to obtain the power deficit / surplus and the power supply-demand difference at each time.

[0011] S5. Statistically analyze the calculation results of all samples to obtain the probability distribution and risk indicators of power imbalance;

[0012] S6. Based on the risk indicators output by the model, construct a three-tiered defense strategy of prevention, emergency response, and recovery.

[0013] Furthermore, the event attribute definition in step S1 of constructing the event-driven model is specifically represented as follows: each event ,in, Indicates the time of occurrence. Indicates spatial location, Indicates the severity. Indicates duration, Indicates the radius of influence.

[0014] Furthermore, the scene classification and association strength quantification in the event-driven model construction in step S1 specifically includes: ① Concurrent scenarios: event time difference The correlation strength is quantified using spatial overlap, specifically expressed as: , This represents a function for calculating the area of ​​a region. Indicates complete overlap. Indicates no overlap; joint impact factor Demonstrates a cumulative effect; ② In-sequence scenario: Events are processed by... The ranking, whose association strength is quantified using state dependency coefficients, is specifically expressed as follows: In the formula, express The system state before the specified time. This indicates that the subsequent event is completely dependent on the state of the preceding event. Indicates independence; ③ Cascading scenario: The event chain meets the triggering conditions. The strength of the association is quantified using propagation probability, specifically expressed as follows: Its value is fitted using historical data. , Indicates the line Power difference Indicates the critical power difference. .

[0015] Furthermore, the construction of the event concurrency scenario model in step S2 specifically includes the following steps: assuming The event in ① The disturbances occur simultaneously within the same timeframe. Calculate the power balance state after their superposition: The power perturbation of each event is Considering spatial overlap The total disturbance is: In the formula, , events , The severity of the situation; ② Considering the reconstruction of the network topology after multiple events cause it to exit, the power balance equation is expressed as: In the formula, Represents a node exist Total output of distributed power sources at any given time. Represents a node exist The charging and discharging power of the energy storage system at all times. Represents a node exist Load power at any given time To exit the route set, The line transmission power is limited by the modified transmission capacity: , For the event For the line The capacity decay coefficient.

[0016] Furthermore, the construction of the event-driven scenario model in step S2 specifically includes the following steps: In the event-driven scenario, events are arranged according to... This involves tracking the temporal evolution of the system state, including: ① constructing a state transition model, when... At that time, the system state vector The update equation is: In the formula, This is the state transition function. For the event The instant after it happened ② Dynamic power balance constraints: Within each event interval, the power balance must satisfy the following: In the formula, For indicator functions, For the event For the region Power disturbance.

[0017] Furthermore, the construction of the event cascading scenario model in step S2 specifically includes the following steps: ① Determine whether the system state exceeds a threshold. Does the event trigger the [number]th event? ① One event; ② After each event, the network topology is corrected using the admittance matrix, and a new power distribution is calculated using DC power flow approximation to achieve power flow reconstruction for fault propagation; ③ The cascaded events terminate when any of the following conditions are met: the probability of the new event Or the system power deficit is less than the reserve capacity. Or, the intersection of the influence radii of all events is empty. .

[0018] Furthermore, in step S3, generating high-dimensional scene samples using vines-Copula specifically includes the following steps: ① For The marginal distributions of the dimensional variables are estimated nonparametrically, specifically using kernel density estimation, whose probability density function is expressed as: In the formula, Indicates the dimension of the variable ( dimension), Indicates bandwidth parameter, express 2. Kernel function; 3. Characterized by the Vines-Copula model. The nonlinear correlation of dimensional variables is decomposed into a series of two-dimensional copulas through a tree structure: In the formula, for Joint Copula function of dimensional variables, For the first The first layer of the tree 1 node The cumulative probability of the marginal distribution. For the first The number of nodes in the layered tree; ③ Using Latin hypercube sampling (LHS) combined with the control variable method, the sampling size is reduced from [previous value] while maintaining sample representativeness. Down to Among them, control variables satisfy And by correcting the estimate To reduce variance, in the formula, For the original estimate, The optimal coefficients are: , The statistic associated with the original estimate.

[0019] Furthermore, the time-segmented solution in step S4 specifically includes a transient phase, a dynamic phase, and a steady-state phase. The transient phase refers to the period from 0 to 1 second after the event occurs, the dynamic phase refers to the period from 1 to 60 seconds after the event occurs, and the steady-state phase refers to the period after the event occurs for more than 60 seconds.

[0020] Furthermore, in step S5, the risk indicator for concurrent event scenarios specifically adopts the expected power deficit. Probabilistic assessment of power balance: In the formula, for Maximum discharge power at all times for The probability density function; the risk indicator for the event-driven scenario in step S5 specifically adopts the conditional value of risk. Assess the cumulative risks associated with the timing of delivery: In the formula, for The power deficit at the quantile level. For confidence level, In step S5, the risk indicator for the event cascading scenario specifically adopts the power transfer distribution factor. This is used to quickly assess the power flow transfer of cascading events, specifically expressed through sensitivity analysis calculations as follows: In the formula, Represents a node Power changes on the line - The impact of transmission power.

[0021] Furthermore, in step S6, the prevention phase optimizes the source-grid-load-storage operation plan for high-risk nm scenarios, including: ① For distributed power sources, limiting wind and solar power to 90% of their rated capacity and reserving a 10% ramp-up margin; ② For energy storage systems, maintaining... To ensure emergency discharge capability; ③ For the power grid network, the voltage is pre-adjusted through on-load tap-changing transformers to improve disturbance rejection margin; In step S6, the emergency phase, based on the real-time prediction results of the model, performs hierarchical control after the event occurs, including: ① Within 0-1s after the event occurs, the maximum power discharge of energy storage is initiated, and spinning reserve is called up; ② Within 1-10s after the event occurs, tertiary loads are disconnected, so that the total load is ≤5%; ③ Within 10-60s after the event occurs, the power flow of the tie line is adjusted, and power is transferred through flexible DC transmission; In step S6, the recovery phase optimizes the recovery sequence after the event subsides: In the formula, For components Recovery time For components Importance weight.

[0022] Due to the adoption of the above technical solutions, the beneficial effects of this invention are: by using an event-driven model to characterize the correlation patterns and transmission rules of multiple events, and at the same time establishing a spatiotemporal probability balance analysis model of public power grid power under different scenarios of concurrent, sequential, and cascaded nm events, and achieving high-dimensional probability compression and balance calculation optimization based on Vines-Copula and importance sampling, it can effectively evaluate the power balance capability and risk of the power grid under different disturbances, and ensure the safe and stable operation of the power system. Attached Figure Description

[0023] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0024] The technical solution of the present invention will be further described in detail below through embodiments and in conjunction with the accompanying drawings.

[0025] like Figure 1 As shown, a method for analyzing the power balance of a public power grid in a nm scenario includes the following analysis steps:

[0026] S1. Based on the event chain and the trigger-propagation-termination rule of time, an event-driven model is constructed. Specifically, it includes event attribute definition, scenario classification and correlation strength quantification. The scenario classification includes concurrent event scenarios, sequential event scenarios and cascading event scenarios.

[0027] Preferably, in step S1, the event attribute definition in constructing the event-driven model is specifically represented as follows: each event ,in, Indicates the time of occurrence. Indicates spatial location, Indicates the severity. Indicates duration, Indicates the radius of influence.

[0028] Preferably, the scene classification and association strength quantification in the event-driven model construction in step S1 specifically includes: ① Concurrent scenarios: event time difference The correlation strength is quantified using spatial overlap, specifically expressed as: , This represents a function for calculating the area of ​​a region. Indicates complete overlap. Indicates no overlap; joint impact factor Demonstrates a cumulative effect; ② In-sequence scenario: Events are processed by... The ranking, whose association strength is quantified using state dependency coefficients, is specifically expressed as follows: In the formula, express The system state before the specified time. This indicates that the subsequent event is completely dependent on the state of the preceding event. Indicates independence; ③ Cascading scenario: The event chain meets the triggering conditions. , such as when the line Exiting caused the line circuit When triggered Transmission; its correlation strength is quantified using transmission probability, specifically expressed as: Its value is fitted using historical data. , Indicates the line Power difference Indicates the critical power difference. Preferably, in cascading scenarios, the evolution of the event chain is described using a Markov state transition matrix: In the formula, express The event state vector at any given time (each element represents the probability of the event occurring). express Transition matrix, ,when ( Can be set to When ), the event chain terminates.

[0029] S2. Construct a spatiotemporal probability balance analysis model for power consumption based on different scenarios, specifically including concurrent event scenario model, sequential event scenario model, and cascading event scenario model.

[0030] Preferably, the construction of the event concurrency scenario model in step S2 specifically includes the following steps: assuming The event in ① The disturbances occur simultaneously within the same timeframe. Calculate the power balance state after their superposition: The power perturbation of each event is Considering spatial overlap The total disturbance is: In the formula, , events , The severity; among them, Specifically, gradient boosting tree (GBDT) can be used to build a data-driven model for prediction. In the formula, For the event exist The input feature vector at each time step can specifically use event type, location, and initial power as input features for this data-driven model; ② Consider the reconstruction of the network topology after multiple events cause it to exit, such as... The power balance equation is then expressed as follows: (The text abruptly ends here, so the translation stops as well.) In the formula, Represents a node exist Total output of distributed power sources at any given time. Represents a node exist The charging and discharging power of the energy storage system at all times. Represents a node exist Load power at any given time To exit the route set, The line transmission power is limited by the modified transmission capacity: , For the event For the line The capacity decay coefficient.

[0031] Preferably, the construction of the event sequential scenario model in step S2 specifically includes the following steps: In the event sequential scenario, events are arranged according to... This involves tracking the temporal evolution of the system state, including: ① constructing a state transition model, when... At that time, the system state vector The update equations (including voltage, power, SOC, etc.) are as follows: In the formula, This is the state transition function. For the event The instant after it happened If the update of the energy storage SOC takes into account the time interval of the spontaneous generation event: In the formula, for ① The charging and discharging efficiency at any given time is affected by temperature and aging; ② Dynamic power balance constraints: within each event interval, the power balance must meet the following requirements: In the formula, For indicator functions, For the event For the region Power disturbance; Decays over time, , is the attenuation constant.

[0032] Preferably, the construction of the event cascading scenario model in step S2 specifically includes the following steps: In the cascading scenario, the propagation of the event chain depends on whether the system state exceeds the critical threshold, and the vicious cycle of "disturbance-response-re-disturbance" needs to be characterized; ① Based on whether the system state exceeds the threshold, determine the first... Does the event trigger the [number]th event? Individual events; such as the triggering condition for line overload: In the formula, For the line The current, For safety margin, it is usually taken as , ① The heat accumulation threshold; ② After each event, the network topology is corrected using the admittance matrix, and a new power distribution is calculated using DC power flow approximation to achieve power flow reconstruction for fault propagation; wherein, the network topology correction using the admittance matrix is ​​expressed as: , The admittance value of the exiting component; the new power distribution calculated using the DC power flow approximation is expressed as: In the formula, Inject power vectors into nodes. For the susceptance matrix, The phase angle vector; ③ The cascade event terminates when any of the following conditions are met: the probability of the newly added event. Or the system power deficit is less than the reserve capacity. Or, the intersection of the influence radii of all events is empty. .

[0033] S3, generated based on event-driven model The association patterns of events are analyzed, and high-dimensional scene samples are generated using vines-Copula. Each sample contains event sequences, power perturbations, and network parameters.

[0034] Preferably, in step S3, generating high-dimensional scene samples using vines-Copula specifically includes the following steps: ① For The marginal distributions of the dimensional variables are estimated nonparametrically, specifically using kernel density estimation, whose probability density function is expressed as: In the formula, Indicates the dimension of the variable ( dimension), Indicates bandwidth parameter, express 2. Kernel function; 3. Characterized by the Vines-Copula model. The nonlinear correlation of dimensional variables is decomposed into a series of two-dimensional copulas through a tree structure: In the formula, for Joint Copula function of dimensional variables, For the first The first layer of the tree 1 node The cumulative probability of the marginal distribution. For the first The number of nodes in the layered tree; selecting the optimal Copula type using the AIC criterion, quantifying the spatiotemporal correlations in different scenarios (such as strong positive correlation in concurrent scenarios and conditional correlation in cascading scenarios); ③ employing Latin hypercube sampling (LHS) combined with the control variable method to maintain sample representativeness while reducing the sampling size from... Down to Among them, control variables satisfy And by correcting the estimate To reduce variance, in the formula, For the original estimate, The optimal coefficients are: , The statistic associated with the original estimate.

[0035] S4. For each sample, solve the power balance equation in different time periods to obtain the power deficit / surplus and the power supply-demand difference at each time.

[0036] Preferably, the time-segmented solution in step S4 specifically includes a transient phase, a dynamic phase, and a steady-state phase, wherein the transient phase refers to the event occurring from 0 to 1 second, the dynamic phase refers to the event occurring from 1 to 60 seconds, and the steady-state phase refers to the event occurring for more than 60 seconds.

[0037] S5. Statistically analyze the calculation results of all samples to obtain the probability distribution and risk indicators of power imbalance.

[0038] Preferably, the risk indicator for concurrent events in step S5 specifically adopts the expected power deficit. Probabilistic assessment of power balance: In the formula, for Maximum discharge power at all times for The probability density function; preferably, when At that time, emergency load shedding measures are triggered.

[0039] Preferably, the risk indicator for the event-driven scenario in step S5 specifically adopts Conditional Value at Risk. Assess the cumulative risks associated with the timing of delivery: In the formula, for The power deficit at the quantile level. For confidence level, By calculating different event sequences Identify high-risk timing patterns, such as multiple events occurring sequentially during peak load periods.

[0040] Preferably, the power transfer distribution factor is specifically used as the risk indicator for the event cascading scenario in step S5. This is used to quickly assess the power flow transfer of cascading events, specifically expressed through sensitivity analysis calculations as follows: In the formula, Represents a node Power changes on the line - The impact of transmission power.

[0041] S6. Based on the risk indicators output by the model, construct a three-tiered defense strategy of prevention, emergency response, and recovery.

[0042] Preferably, the prevention phase in step S6 optimizes the source-grid-load-storage operation plan for high-risk nm scenarios, including: ① For distributed power sources, limiting wind and solar power to 90% of their rated capacity and reserving a 10% ramp-up margin; ② For energy storage systems, maintaining... ③ Ensure emergency discharge capability; ③ For the power grid network, improve the disturbance immunity margin by pre-adjusting the voltage through on-load tap-changing transformers.

[0043] Preferably, in step S6, the emergency phase, based on the real-time prediction results of the model after the event occurs, performs hierarchical control, including: ① within 0-1 seconds after the event occurs, starting the maximum power discharge of the energy storage and calling up the spinning standby (such as gas turbines); ② within 1-10 seconds after the event occurs, cutting off tertiary loads (such as non-critical industrial loads) so that the total load is ≤5%; ③ within 10-60 seconds after the event occurs, adjusting the power flow of the tie line and transferring power through flexible DC transmission.

[0044] Preferably, in step S6, the recovery phase optimizes the recovery sequence after the event has subsided: In the formula, For components Recovery time For components Importance weight (load center weight is 1.2).

[0045] Preferably, taking a 220kV grid of a provincial power grid as the background, two n-2 scenarios are simulated, including concurrent line tripping and cascaded line tripping; wherein, ① in the concurrent scenario, the scenario parameters are set to two tie lines (L1, L2) trip simultaneously at t=0s, each line transmits 150MW of power, and the spatial overlap is... The model output results are: the total power deficit instantaneously reaches 280MW (considering the superposition effect), and the energy storage system outputs 100MW within 0.2s. At a 95% confidence level ② In a cascaded scenario, the scenario parameters are set so that L1 trips at t=0s, causing L3 to overload (current reaches 1.2). The L3 circuit breaker trips at t=0.5s, with a conduction probability of p=0.9. The model output is: an initial deficit of 150MW, and a secondary deficit that adds up to 270MW. Because the energy storage system has already consumed some capacity in the first event, the secondary response capability drops to 60MW. , Therefore, compared with the actual scheduling data, the peak error of the power deficit is <5%. With an error of less than 8%, it can effectively characterize the coupled effects of multiple events.

[0046] 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 scope of the claims of the present invention.

Claims

1. A method for analyzing the power balance of a public power grid in a nm scenario, characterized in that: The analysis includes the following steps: S1. Based on the event chain as the core and the trigger-propagation-termination rule of time, an event-driven model is constructed, which includes event attribute definition, scenario classification and correlation strength quantification. The scenario classification includes concurrent event scenarios, sequential event scenarios and cascading event scenarios. S2. Construct a spatiotemporal probability balance analysis model for power consumption based on different scenarios, specifically including event concurrency scenario model, event sequential scenario model, and event cascading scenario model; S3, generated based on event-driven model The correlation patterns of events are analyzed, and high-dimensional scene samples are generated using vines-Copula. Each sample contains event sequences, power perturbations, and network parameters. S4. For each sample, solve the power balance equation in different time periods to obtain the power deficit / surplus and the power supply-demand difference at each time. S5. Statistically analyze the calculation results of all samples to obtain the probability distribution and risk indicators of power imbalance; S6. Based on the risk indicators output by the model, construct a three-tiered defense strategy of prevention, emergency response, and recovery.

2. The method for analyzing the power balance of a public power grid in a nm scenario according to claim 1, characterized in that: In step S1, the event attribute definition in constructing the event-driven model is specifically represented as follows: Each event ,in, Indicates the time of occurrence. Indicates spatial location, Indicates the severity. Indicates duration, Indicates the radius of influence.

3. The method for analyzing the power balance of a public power grid in a nm scenario according to claim 1, characterized in that: The scenario classification and correlation strength quantification in the event-driven model construction step S1 specifically includes: ① Concurrent scenarios: event time difference The correlation strength is quantified using spatial overlap, specifically expressed as: , This represents a function for calculating the area of ​​a region. Indicates complete overlap. Indicates no overlap; joint impact factor Demonstrates a cumulative effect; ② In-sequence scenario: Events are processed by... The ranking, whose association strength is quantified using state dependency coefficients, is specifically expressed as follows: In the formula, express The system state before the specified time. This indicates that the subsequent event is completely dependent on the state of the preceding event. Indicates independence; ③ Cascading scenario: The event chain meets the triggering conditions. The strength of the association is quantified using propagation probability, specifically expressed as follows: Its value is fitted using historical data. , Indicates the line Power difference Indicates the critical power difference. .

4. The method for analyzing the power balance of a public power grid in a nm scenario according to claim 1, characterized in that: The construction of the event concurrency scenario model in step S2 specifically includes the following steps: assuming The event in ① The disturbances occur simultaneously within the same timeframe. Calculate the power balance state after their superposition: The power perturbation of each event is Considering spatial overlap The total disturbance is: In the formula, , events , The severity of the situation; ② Considering the reconstruction of the network topology after multiple events cause it to exit, the power balance equation is expressed as: In the formula, Represents a node exist Total output of distributed power sources at any given time. Represents a node exist The charging and discharging power of the energy storage system at all times. Represents a node exist Load power at any given time To exit the route set, The line transmission power is limited by the modified transmission capacity: , For the event For the line The capacity decay coefficient.

5. The method for analyzing the power balance of a public power grid in a nm scenario according to claim 1, characterized in that: The construction of the event-driven scenario model in step S2 specifically includes the following steps: In the event-driven scenario, events are arranged according to... This involves tracking the temporal evolution of the system state, including: ① constructing a state transition model, when... At that time, the system state vector The update equation is: In the formula, This is the state transition function. For the event The instant after it happened ② Dynamic power balance constraints: Within each event interval, the power balance must satisfy the following: In the formula, For indicator functions, For the event For the region Power disturbance.

6. The method for analyzing the power balance of a public power grid in a nm scenario according to claim 1, characterized in that: The construction of the event cascading scenario model in step S2 specifically includes the following steps: ① Determine whether the system state exceeds the threshold. Does the event trigger the [number]th event? ① One event; ② After each event, the network topology is corrected using the admittance matrix, and a new power distribution is calculated using DC power flow approximation to achieve power flow reconstruction for fault propagation; ③ The cascaded events terminate when any of the following conditions are met: the probability of the new event Or the system power deficit is less than the reserve capacity. Or, the intersection of the influence radii of all events is empty. .

7. The method for analyzing the power balance of a public power grid in a nm scenario according to claim 1, characterized in that: Step S3, which generates high-dimensional scene samples using vines-Copula, specifically includes the following steps: ① For The marginal distributions of the dimensional variables are estimated nonparametrically, specifically using kernel density estimation, whose probability density function is expressed as: In the formula, Indicates the dimension of the variable ( dimension), Indicates bandwidth parameter, express 2. Kernel function; 3. Characterized by the Vines-Copula model. The nonlinear correlation of dimensional variables is decomposed into a series of two-dimensional copulas through a tree structure: In the formula, for Joint Copula function of dimensional variables, For the first The first layer of the tree 1 node The cumulative probability of the marginal distribution. For the first The number of nodes in the layered tree; ③ Using Latin hypercube sampling (LHS) combined with the control variable method, the sampling size is reduced from [previous value] while maintaining sample representativeness. Down to Among them, control variables satisfy And by correcting the estimate To reduce variance, in the formula, For the original estimate, The optimal coefficients are: , The statistic associated with the original estimate.

8. The method for analyzing the power balance of a public power grid in a nm scenario according to claim 1, characterized in that: The time-segmented solution in step S4 specifically includes a transient phase, a dynamic phase, and a steady-state phase. The transient phase refers to the period from 0 to 1 second after the event occurs, the dynamic phase refers to the period from 1 to 60 seconds after the event occurs, and the steady-state phase refers to the period after the event occurs for more than 60 seconds.

9. The method for analyzing the power balance of a public power grid in a nm scenario according to claim 1, characterized in that: In step S5, the risk indicator for concurrent events specifically adopts the expected power deficit. Probabilistic assessment of power balance: In the formula, for Maximum discharge power at all times for The probability density function; the risk indicator for the event-driven scenario in step S5 specifically adopts the conditional value of risk. Assess the cumulative risks associated with the timing of delivery: In the formula, for The power deficit at the quantile level. For confidence level, In step S5, the risk indicator for the event cascading scenario specifically adopts the power transfer distribution factor. This is used to quickly assess the power flow transfer of cascading events, specifically expressed through sensitivity analysis calculations as follows: In the formula, Represents a node Power changes on the line - The impact of transmission power.

10. The method for analyzing the power balance of a public power grid in a nm scenario according to claim 1, characterized in that: The prevention phase in step S6 optimizes the source-grid-load-storage operation plan for high-risk nm scenarios, including: ① For distributed power sources, limiting wind and solar power to 90% of their rated capacity and reserving a 10% ramp-up margin; ② For energy storage systems, maintaining... To ensure emergency discharge capability; ③ For the power grid network, the voltage is pre-adjusted through on-load tap-changing transformers to improve disturbance rejection margin; In step S6, the emergency phase, based on the real-time prediction results of the model, performs hierarchical control after the event occurs, including: ① Within 0-1s after the event occurs, the maximum power discharge of energy storage is initiated, and spinning reserve is called up; ② Within 1-10s after the event occurs, tertiary loads are disconnected, so that the total load is ≤5%; ③ Within 10-60s after the event occurs, the power flow of the tie line is adjusted, and power is transferred through flexible DC transmission; In step S6, the recovery phase optimizes the recovery sequence after the event subsides: In the formula, For components Recovery time For components Importance weight.

Images