A hybrid optimization method and system for island-type heterogeneous source-load-storage inertia assessment

Abstract

The present invention relates to a hybrid optimization method and system for island-type heterogeneous source load-storage inertia assessment. This method first models multi-source inertia and constructs a spatiotemporal distribution discrete model based on the multi-source inertia dynamic model and related parameters. An ant colony algorithm is then used to generate an initial solution set for the spatiotemporal distribution discrete model. A determination is made as to whether the solution set satisfies a first constraint; if not, the solution set is returned and regenerated. Finally, a genetic algorithm is used to further process the solution set, output an optimized solution set, and determine whether the optimized solution set satisfies a second constraint. If so, the optimal inertia configuration is output; if not, the optimization is returned and continued. Compared with the prior art, the present invention has advantages such as improving the effectiveness of inertia optimization results.

Description

Technical Field

[0001] The present invention relates to the technical field of power optimization, and in particular to an island-type heterogeneous source-load-storage inertia assessment hybrid optimization method and system. Background Art

[0002] The global energy mix is ​​rapidly transitioning toward renewable energy. Wind power and photovoltaics are rapidly expanding, accounting for over 50% of total installed capacity, marking a new era in which the power system is dominated by renewable energy. This transformation is accompanied by the continued advancement of power electronics across the entire "source-grid-load" chain, fostering a new power system architecture characterized by the deep integration of high renewable energy penetration and high-density power electronics. Microgrid systems, a typical application scenario, typically feature distributed power sources such as wind power, photovoltaics, and energy storage, accounting for over 70% of their power mix. These devices are connected to the grid through power electronic converters. While offering the advantages of flexible regulation, they lack the mechanical inertia and electromagnetic damping characteristics of synchronous generators. Research has shown that when system inertia levels fall below a critical threshold, the frequency dynamics of microgrids under sudden load changes or fault conditions exhibit increased overshoot and prolonged stabilization times. In extreme cases, this can lead to cascading grid disconnections. This degraded dynamic stability, caused by a lack of physical inertia, has become a core bottleneck hindering the safe operation of new power systems.

[0003] The traditional inertia assessment system based on the synchronous machine-dominated architecture is difficult to effectively characterize the isolated microgrid system with the dynamic coupling characteristics of heterogeneous sources, loads and storage. When the proportion of fluctuating power sources such as wind power / photovoltaic power exceeds 65% and the load exhibits multi-time scale fluctuation characteristics, the interaction between the maximum power point tracking control on the source side and the response mechanism of the converter on the storage side will cause the system equivalent inertia to exhibit time-varying nonlinear characteristics. Existing research has not yet established an inertia demand quantification model that considers the dynamic matching degree of heterogeneous units. In particular, under multi-dimensional random perturbations of sources, loads and storage, how to construct the minimum inertia boundary criterion that takes into account the voltage / frequency coupling effect has become a core problem to ensure the dynamic stability of the isolated microgrid. In addition, traditional algorithms often use linear approximation or fixed proportion method optimization, which is prone to falling into local optimality.

[0004] For example, the invention patent with publication number CN117767351A only controls virtual synchronous generator power systems. When addressing the interaction between source-side fluctuating power sources (such as wind power and photovoltaics) and the response mechanisms of the load-storage converter, it may not fully account for the complex dynamic coupling characteristics under multi-dimensional random perturbations. For example, in actual operation, when multiple factors such as sudden changes in wind speed and light intensity, as well as rapid changes in load, act simultaneously, it is difficult to accurately characterize the time-varying nonlinear characteristics of the system's equivalent inertia. This method uses a simulated annealing algorithm to adjust virtual inertia regulation and damping, but this algorithm suffers from slow convergence. In large-scale systems or scenarios with high real-time requirements, it may not be possible to quickly obtain the optimal target virtual inertia regulation coefficient and damping. Furthermore, the simulated annealing algorithm relies on an initial solution. If the initial solution is not chosen properly, it may become trapped in a local optimum and fail to find the true global optimal solution.

[0005] In summary, traditional inertia evaluation oversimplifies the system dynamic characteristics model and fails to effectively characterize the coupling and regulation characteristics of various types of new energy sources, resulting in problems such as excessive quantization errors. This leads to problems such as traditional inertia optimization being unable to adapt to the new power system structure, lacking dynamic matching considerations, and ignoring the coupling effects of multiple factors. As a result, the inertia optimization results are not accurate and effective enough, and the effect is not good in practical applications, especially in scenarios with high new energy penetration rates, where traditional inertia calculation errors are large. Summary of the Invention

[0006] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and provide an island-type heterogeneous source-load-storage inertia assessment hybrid optimization method and system.

[0007] The purpose of the present invention can be achieved by the following technical solutions:

[0008] According to one aspect of the present invention, a hybrid optimization method for evaluating inertia of isolated heterogeneous source-load-storage is provided, the method comprising the following steps:

[0009] S1. Model the multi-source inertia and construct a spatiotemporal distribution discrete model based on the multi-source inertia dynamic model and related parameters;

[0010] S2, using ant colony algorithm to generate the initial solution set of spatiotemporal distribution discrete model;

[0011] S3. Determine whether the solution set satisfies the first constraint. If so, execute step S4; otherwise, return to execute step S2;

[0012] S4, further processing the solution set using genetic algorithm and outputting the optimized solution set;

[0013] S5. Determine whether the optimized solution set satisfies the second constraint. If so, output the optimal inertia configuration; otherwise, return to step S4.

[0014] As the preferred technical solution, the specific process of multi-source inertia modeling in S1 is as follows: first, discrete state equations for the new energy units using VSG control are constructed separately, and then dual-mode discrete control of the energy storage system is introduced; the specific formula is:

[0015] ;

[0016] ;

[0017] in, For the i The virtual inertia time constant of typhoon power; For the i The virtual inertia time constant of the photovoltaic system; For the k Output power of wind turbine at any moment; For the k Output power of the photovoltaic unit at any moment; It is the virtual inertia of new energy; N is the total number of new energy units; is the discretized time interval; For the energy storage system k The virtual inertia time constant at each discrete moment; For the k The change in system frequency at a discrete moment; is the frequency change rate threshold; is the virtual inertia time constant when the energy storage system is in inertia control mode; is the virtual inertia time constant when the energy storage system is in droop control mode.

[0018] As a preferred technical solution, the spatiotemporal distribution discrete model constructed in S1 includes a system frequency change rate model and a power disturbance model. The system frequency change rate model is used to describe the rate of change of the power system frequency over time, reflecting the dynamic response characteristics of the system frequency when affected by various factors; the power disturbance model is used to characterize the power imbalance in the power system and study the various power disturbance factors affecting the system and their impacts. The specific formula is:

[0019] ;

[0020] in, For the system k Frequency change at each moment; is the discretized time interval; is the system frequency variation; for k Total system inertia requirement at any moment; for k The power imbalance caused by power disturbances in the system at any moment, such as sudden load changes and power generation equipment failures; for k The power response of the system at that moment, that is, the power adjustment amount made by the system in response to the power disturbance; for k Virtual power response at each moment; for k The power response quantity related to the unit inertia at any time.

[0021] As a preferred technical solution, the spatiotemporal distribution discrete model is modified in real time based on the difference between the unit frequency and the inertia center frequency, according to the inertia distribution coefficient. If the difference between the unit frequency and the inertia center frequency is large, it indicates that the unit inertia is unevenly distributed, and the total system inertia demand and the power response related to the unit inertia need to be adjusted. If the distribution coefficient approaches 1, it indicates that the unit inertia is evenly distributed and no adjustment is required. The specific formula for the inertia distribution coefficient is:

[0022] ;

[0023] in, is the system frequency based on the center of inertia; for k moment Inertia distribution coefficient; N is the total number of units in the system; For the i Units in k The frequency of the moment.

[0024] As a preferred technical solution, k The power response quantity related to the unit inertia at the moment is obtained by k The load power at each moment is corrected and the specific formula is:

[0025] ;

[0026] in, is the load voltage equivalent inertia correction; is the reference load power; for k The voltage value of the system at the moment, is the reference voltage value; is constant impedance; is the constant current load ratio.

[0027] As a preferred technical solution, the first constraint in S3 is the RoCoF constraint, and its specific formula is:

[0028] ;

[0029] in, is the value of the system frequency change rate based on the inertia center at time k, reflecting how fast the frequency changes with time; The maximum frequency change rate threshold allowed is used to limit the frequency change speed and ensure system stability.

[0030] As a preferred technical solution, the specific process of further processing the solution set using the genetic algorithm in S4 is: using the penalty function to process the constraints and then solve the optimal solution. The specific formula is:

[0031] ;

[0032] in, For k The penalty function value at the moment is used to measure the degree to which the solution violates the constraints; α is the penalty coefficient, which is used to adjust the sensitivity of the penalty function to constraint violations; is the rate of change of system frequency based on the center of inertia k The value of the moment; The maximum frequency change rate threshold allowed.

[0033] As a preferred technical solution, the second constraint in S5 includes a frequency minimum point constraint and an energy storage frequency capacity constraint; wherein, the frequency minimum point constraint is that the actual value of the system frequency minimum point is greater than or equal to the minimum frequency value allowed by the system; the energy storage frequency capacity constraint is that k The power change of the energy storage system participating in frequency regulation at any moment is less than or equal to the maximum frequency regulation power capacity allowed by the energy storage system. The specific formula is:

[0034] ;

[0035] ;

[0036] in, is the actual value of the lowest frequency point of the system; is the lowest frequency value allowed by the system; For k The power change of the energy storage system participating in frequency regulation at each moment; Adjust the power capacity for the maximum frequency allowed by the energy storage system.

[0037] As a preferred technical solution, the actual value of the lowest frequency point of the system is obtained by the sliding window extreme value detection method. First, the time length of the sliding window is set; then the time of the lowest frequency point is determined based on the time length of the sliding window; finally, the frequency value at that moment is used to determine the lowest frequency point. Calculate the actual lowest frequency , the specific formula is:

[0038] ;

[0039] in, is the moment of lowest frequency; is the system frequency based on the center of inertia; k is the index value of the discrete time series; is the index value corresponding to the starting time of the sliding window; T is the length of the sliding window.

[0040] According to another aspect of the present invention, a hybrid optimization system for evaluating inertia of an isolated heterogeneous source, load, and storage is provided, the system comprising a model building module, a solution set generation and preliminary processing module, and a solution set optimization and judgment module;

[0041] The model building module is used to model multi-source inertia, obtain and construct a spatiotemporal distribution discrete model based on the multi-source inertia dynamic model and related parameters;

[0042] The solution set generation and preliminary processing module is used to generate the initial solution set of the spatiotemporal distribution discrete model using the ant colony algorithm, and determine whether the solution set meets the first constraint. If not, it returns to regenerate;

[0043] The solution set optimization and judgment module is used to further process the solution set using the genetic algorithm, output the optimized solution set, and judge whether the optimized solution set meets the second constraint. If it meets the constraint, the optimal inertia configuration is output; if not, the module returns to continue optimization.

[0044] Compared with the prior art, the present invention has the following beneficial effects:

[0045] 1. The present invention first models multi-source inertia and constructs a spatiotemporal distribution discrete model based on the multi-source inertia dynamic model and related parameters. Regarding the source of inertia, this method integrates new energy virtual inertia, dynamic regulation of energy storage, and load demand response to construct a multi-source collaborative inertia system, fully tapping the frequency regulation potential of distributed resources. Furthermore, in terms of spatiotemporal characteristics, this method solves the problem of traditional methods ignoring the distribution differences of sources, loads, and storage in the island mode, resulting in evaluation and optimization results deviating from actual needs. This method constructs a spatiotemporal distribution discrete model, quantifies the frequency difference between the disturbance point and the inertia center, and dynamically corrects the local inertia weight, significantly improving the evaluation accuracy and thus the effectiveness of the inertia optimization results. In particular, in scenarios with high new energy penetration, the inertia calculation error of the proposed method is greatly reduced.

[0046] 2. The present invention can dynamically adapt to changes in the unit's inertia distribution by correcting the spatiotemporal distribution discrete model in real time based on the inertia distribution coefficient. The power self-correction model based on discrete voltage sampling solves the problem of inflated inertia caused by static load voltage fluctuations, thereby improving the accuracy of the model's description of the system's inertia characteristics.

[0047] 3. The present invention obtains the power response related to the unit inertia by correcting the load power, taking into account the influence of factors such as voltage, so that the power response calculation is more in line with the actual operating conditions, enhancing the accuracy of system analysis, and thus improving the accuracy of the optimization results.

[0048] 4. This invention proposes an ACO-GA hybrid optimization strategy by employing a hybrid optimization strategy, combining the global optimization capabilities of a genetic algorithm and an ant colony algorithm. This strategy utilizes the ant colony algorithm (ACO) to rapidly generate an initial set of solutions satisfying the RoCoF constraints in the discrete solution space. The dynamic penalty function mechanism of the genetic algorithm (GA) then refines the frequency minimum point constraint, achieving efficient coordination between global and local searches. Furthermore, by combining the primary and secondary constraints, this strategy efficiently solves multi-constrained nonlinear models, enabling flexible adjustment of search directions under complex constraints. This improves the algorithm's constraint handling capabilities and helps it approach the global optimal solution. Furthermore, this strategy reduces parameter sensitivity, ensuring both economical and safe inertia configuration.

[0049] 5. This invention breaks through the limitations of traditional methods, which are often limited to a single indicator (such as RoCoF or frequency extremes), integrates the multiple constraints of RoCoF, frequency extremes, and spare capacity, and considers the response speed and capacity limitations of frequency regulation resources to form an evaluation framework that is more in line with actual operations. By setting RoCoF constraints, the frequency change rate is limited, providing a quantitative guarantee standard for system stability, effectively preventing system instability caused by excessive frequency changes; and dynamically switching between inertia and droop modes based on the RoCoF threshold, taking into account both rapid response and frequency regulation economy. By setting the minimum frequency point and energy storage frequency capacity constraints, the stable operation of the system is guaranteed from the perspectives of frequency extremes and energy storage regulation capabilities, ensuring the system's frequency safety under different operating conditions.

[0050] 6. In the present invention, the actual value of the lowest point of the system frequency is obtained by adopting the sliding window extreme value detection method, which can effectively capture the frequency extreme value moment and value in real time, providing key data support for system stability evaluation. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 Schematic diagram of the steps of a hybrid optimization method for evaluating inertia of isolated-island heterogeneous source-load-storage in the present invention;

[0052] Figure 2 A schematic diagram of a process for performing hybrid optimization of inertia evaluation in an embodiment;

[0053] Figure 3 Schematic diagram of the process of generating a solution set using the ant colony optimization algorithm in an embodiment;

[0054] Figure 4 Schematic diagram of the process of optimizing the solution set using the genetic algorithm in the embodiment. DETAILED DESCRIPTION

[0055] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0056] This application proposes a hybrid optimization method and system for inertia assessment of isolated heterogeneous sources, loads and storage. The proposed method can be applied to the operation scenarios of isolated microgrids containing a large amount of renewable energy. In order to solve the problems that traditional inertia assessment oversimplifies the dynamic characteristics of the system and fails to effectively characterize the coupling regulation characteristics of various renewable energy sources, resulting in excessive quantization errors, a collaborative operation topology of multiple types of new energy equipment is constructed, the spatiotemporal response characteristics of sources, loads and storage are accurately quantified, and a unified inertia assessment method for heterogeneous sources, loads and storage is proposed to accurately assess the dynamic inertia level of the system.

[0057] Example 1

[0058] In this embodiment, with frequency stability in island mode as the core goal, a closed-loop evaluation process of multi-source modeling - spatiotemporal correction - optimization solution - simulation verification is constructed;

[0059] In this embodiment, a hybrid optimization method for evaluating the inertia of an isolated heterogeneous source load storage is applied. The method steps are as follows: Figure 1 As shown, specifically including:

[0060] S1. Model the multi-source inertia and construct a spatiotemporal distribution discrete model based on the multi-source inertia dynamic model and related parameters;

[0061] S2, using ant colony algorithm to generate the initial solution set of spatiotemporal distribution discrete model;

[0062] S3. Determine whether the solution set satisfies the first constraint. If so, execute step S4; otherwise, return to execute step S2;

[0063] S4, further processing the solution set using genetic algorithm and outputting the optimized solution set;

[0064] S5. Determine whether the optimized solution set satisfies the second constraint. If so, output the optimal inertia configuration; otherwise, return to step S4.

[0065] Specific implementation such as Figure 2 The process shown includes the following:

[0066] 1. First, perform modeling and parameter calculation of the discretized multi-source inertia dynamics;

[0067] The discrete state equation of the new energy unit (wind power / photovoltaic) using VSG control is:

[0068] ;

[0069] in, For the i The virtual inertia time constant of typhoon power; For the i The virtual inertia time constant of the photovoltaic system; For the k Output power of wind turbine at any moment; For the k Output power of the photovoltaic unit at any moment; It is the virtual inertia of new energy; N is the total number of new energy units; is the discretized time interval.

[0070] Energy storage system introduces dual-mode discrete control:

[0071] ;

[0072] in, For the energy storage system k The virtual inertia time constant at each discrete moment; For the k The change in system frequency at a discrete moment; is the discretized time interval; is the frequency change rate threshold; is the virtual inertia time constant when the energy storage system is in inertia control mode; is the virtual inertia time constant when the energy storage system is in droop control mode.

[0073] 2. Then build a discrete model of spatiotemporal distribution

[0074] The discrete state equation of the discrete model is:

[0075]

[0076] in, For the system k Frequency change at each moment; is the discretized time interval; is the system frequency variation; for k Total system inertia requirement at any moment; for k The power imbalance caused by power disturbances in the system at any moment, such as sudden load changes and power generation equipment failures; for k The power response of the system at that moment, that is, the power adjustment amount made by the system in response to the power disturbance; for k Virtual power response at each moment; for k The power response quantity related to the unit inertia at any time.

[0077] Correct the load power at the kth moment:

[0078] ;

[0079] in, is the load voltage equivalent inertia correction; is the reference load power; for k The voltage value of the system at the moment, is the reference voltage value; is constant impedance; is the constant current load ratio.

[0080] Update of inertia distribution coefficient:

[0081] ;

[0082] in, is the system frequency based on the center of inertia; for k moment Inertia distribution coefficient; N is the total number of units in the system; For the i Units in k The frequency of the moment.

[0083] Sliding window extreme value detection:

[0084] ;

[0085] in, is the moment of lowest frequency; is the system frequency based on the center of inertia; k is the index value of the discrete time series; is the index value corresponding to the starting time of the sliding window; T is the length of the sliding window.

[0086] 3. Then build a constrained hybrid optimization solution model;

[0087] Among them, the objective function is:

[0088] ;

[0089] in, for k Total system inertia requirement at any moment; Mis the total number of time steps considered during the optimization process; k is the discrete time step index.

[0090] The constraints include:

[0091] RoCoF constraints:

[0092] ;

[0093] in, is the value of the system frequency change rate based on the inertia center at time k, reflecting how fast the frequency changes with time; The maximum frequency change rate threshold allowed is used to limit the frequency change speed and ensure system stability.

[0094] Frequency minimum:

[0095] ;

[0096] in, is the actual value of the lowest frequency point of the system; The lowest frequency value allowed by the system.

[0097] Energy storage frequency capacity:

[0098] ;

[0099] in, For k The power change of the energy storage system participating in frequency regulation at each moment; Adjust the power capacity for the maximum frequency allowed by the energy storage system.

[0100] The hybrid optimization algorithms used include:

[0101] Ant Colony Optimization Algorithm (ACO): Generates discrete solution sets , quickly search the feasible domain, where, is the total inertia of the system; the discrete solution set generated by the ant colony algorithm, where Each element in Represents a possible configuration of the total inertia of the system.

[0102] Genetic Algorithm (GA): uses penalty functions to handle constraints and accurately find the optimal solution

[0103] ;

[0104] in, For kThe penalty function value at the moment is used to measure the degree to which the solution violates the constraints; α is the penalty coefficient, which is used to adjust the sensitivity of the penalty function to constraint violations; is the rate of change of system frequency based on the center of inertia k The value of the moment; The maximum frequency change rate threshold allowed.

[0105] Finally, output the optimal inertia configuration requirements .

[0106] Example 2

[0107] In this embodiment, an island-type heterogeneous source-load-storage inertia assessment hybrid optimization system is adopted, which includes a model construction module, a solution set generation and preliminary processing module, and a solution set optimization and judgment module;

[0108] The model building module is used to model multi-source inertia, obtain and construct a spatiotemporal distribution discrete model based on the multi-source inertia dynamic model and related parameters;

[0109] The solution set generation and preliminary processing module is used to generate the initial solution set of the spatiotemporal distribution discrete model using the ant colony algorithm, and determine whether the solution set meets the first constraint. If not, it returns to regenerate;

[0110] The solution set optimization and judgment module is used to further process the solution set using the genetic algorithm, output the optimized solution set, and judge whether the optimized solution set meets the second constraint. If it meets the constraint, the optimal inertia configuration is output; if not, the module returns to continue optimization.

[0111] The system works by adopting a hybrid optimization method for evaluating inertia of isolated heterogeneous source-load-storage as described in Example 1;

[0112] In this embodiment, the solution set generation and preliminary processing module uses the heuristic search characteristics of the ant colony algorithm to find the solution set in the discretized system total inertia configuration space.

[0113] ;

[0114] The algorithm quickly generates an initial set of feasible solutions that satisfy the RoCoF constraints. The algorithm simulates the foraging behavior of ants and uses the pheromone update mechanism to guide the search direction, giving priority to exploring areas with high feasible solutions.

[0115] The workflow of solution generation and preliminary processing module is as follows Figure 3 As shown, it specifically includes:

[0116] 1. Initialization parameters

[0117] First set the number of ants m , pheromone heuristic factors α, Expectation Heuristic Factor β , pheromone evaporation rate ρ and the maximum number of iterations .

[0118] Define the feasible region of total inertia of the system , and discretized into N candidate values:

[0119] .

[0120] Initialize the pheromone matrix to represent the initial attraction of each candidate inertia configuration.

[0121] 2. Ant traversal search

[0122] Each ant selects an inertia configuration from the candidate inertia set. ;

[0123] 3. Constraint Verification (RoCoF Constraints)

[0124] Inertia configuration generated for each ant , calculate the frequency change rate based on the center of inertia; determine whether the RoCoF constraint is met;

[0125] If the constraints are met, add the configuration to the initial solution set and record its objective function value (the sum of the total inertia requirements of the system);

[0126] 4. Pheromone Update

[0127] For feasible solutions , update the pheromone according to the following formula:

[0128] ;

[0129] ;

[0130] in, is the path at time t ( i , j ) on the pheromone concentration; For the k The amount of pheromones released by ants.

[0131] 5. Termination condition judgment

[0132] If the maximum number of iterations is reached Or if the size of the initial solution set reaches the target, the search is terminated and the solution set is output; otherwise, return 2 and continue searching.

[0133] In this embodiment, the solution set optimization and judgment module uses the global search capability of the genetic algorithm to further optimize the initial feasible solution set generated by the ant colony algorithm. Through selection, crossover, mutation operations and penalty function processing constraints, it gradually approaches the optimal inertia configuration that meets the frequency minimum point constraint and the energy storage frequency capacity constraint.

[0134] In this embodiment, the workflow of the solution set optimization and judgment module is as follows: Figure 4 As shown, it specifically includes:

[0135] 1. Population initialization

[0136] Initialize the solution set and use the initial solution set as the initial population of the genetic algorithm , individual is inertia configuration .

[0137] 2. Fitness function calculation

[0138] The fitness function is defined as:

[0139] ;

[0140] in, is the penalty function; is the total number of time steps considered during the optimization process.

[0141] 3. Select an operation

[0142] The roulette wheel selection method is used to select individuals to enter the next generation population according to their fitness values. Individuals with higher fitness have a greater probability of being selected.

[0143] 4. Crossover Operation

[0144] The crossover probability of the selected individual Perform a single-point crossover to generate new individuals. For example, randomly select a crossover point and exchange some genes (inertia configuration parameters) of the two parent individuals.

[0145] 5. Mutation Operation

[0146] For new individuals, the mutation probability Perform mutation, randomly adjust its inertia configuration parameters (such as randomly jumping among discrete candidate values), and introduce new search directions.

[0147] 6. Constraint Verification

[0148] For the mutated individuals, the frequency minimum point constraint and energy storage frequency capacity constraint are checked. If the constraints are met, the individual is retained; otherwise, its fitness is adjusted using a penalty function or it is directly eliminated.

[0149] 7. Termination condition judgment

[0150] If the maximum number of iterations is reached Or the fitness value converges, terminate the algorithm and output the optimal solution set; otherwise return to 3 and continue iterating.

[0151] In summary, this solution proposes to incorporate new energy virtual inertia, energy storage dynamic response and load voltage characteristics into a unified modeling framework, breaking through the limitation of traditional methods that rely solely on the inertia of synchronous units. Based on the power self-correction model of discrete voltage sampling, the problem of high inertia caused by static load voltage fluctuation is solved, and the inertia / droop mode is dynamically switched according to the RoCoF threshold, taking into account both fast response and frequency regulation economy. In scenarios with high new energy penetration, the inertia calculation error is greatly reduced.

[0152] This solution innovatively embeds the inertia distribution coefficient by improving the model quantifying differences in frequency spatial distribution, this solves the problem of local frequency prediction bias caused by traditional models relying solely on the center frequency of inertia. Furthermore, an ACO-GA hybrid optimization strategy is proposed, which utilizes an ant colony algorithm to rapidly generate an initial solution set that satisfies the RoCoF constraint in the discrete solution space. The dynamic penalty function mechanism of the genetic algorithm then refines the frequency minimum point constraint, achieving efficient coordination between global and local searches.

[0153] This solution pioneers a fully discrete modeling framework, employing discrete representations for everything from power sampling and state equations to optimization variables, seamlessly adapting to digital simulation and hardware deployment requirements. Based on this, a closed-loop control strategy, "storage first, synchronous units in reserve," is proposed. This prioritizes the virtual inertia of energy storage to quickly compensate for shortfalls, and synchronous units are started and stopped only in extreme scenarios, balancing economic efficiency and reliability.

[0154] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and such modifications or substitutions are intended to be within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be subject to the scope of protection of the claims.

Claims

1. A hybrid optimization method for island-type heterogeneous source-load-storage inertia assessment, characterized by: The method steps include: S1. Model the multi-source inertia and construct a spatiotemporal distribution discrete model based on the multi-source inertia dynamic model and related parameters; S2, using ant colony algorithm to generate the initial solution set of spatiotemporal distribution discrete model; S3. Determine whether the solution set satisfies the first constraint. If so, execute step S4; otherwise, return to execute step S2; S4, further processing the solution set using genetic algorithm and outputting the optimized solution set; S5. Determine whether the optimized solution set satisfies the second constraint. If so, output the optimal inertia configuration; otherwise, return to step S4. The specific process of multi-source inertia modeling in S1 is as follows: first, discrete state equations of the new energy units using VSG control are constructed, and then the dual-mode discrete control of the energy storage system is introduced; the specific formula is: ； ； in, For the i The virtual inertia time constant of typhoon power; For the i The virtual inertia time constant of the photovoltaic system; For the k Output power of wind turbine at any moment; For the k Output power of the photovoltaic unit at any moment; It is the virtual inertia of new energy; N is the total number of new energy units; is the discretized time interval; For the energy storage system k The virtual inertia time constant at each discrete moment; For the k The change in system frequency at a discrete moment; is the frequency change rate threshold; is the virtual inertia time constant when the energy storage system is in inertia control mode; is the virtual inertia time constant when the energy storage system is in droop control mode; The spatiotemporal distribution discrete model constructed in S1 includes a system frequency change rate model and a power disturbance model. The system frequency change rate model is used to describe the rate of change of the power system frequency over time, reflecting the dynamic response characteristics of the system frequency when affected by various factors; the power disturbance model is used to characterize the power imbalance in the power system and study the various power disturbance factors affecting the system and their impacts. The specific formula is: ； in, For the system k Frequency change at each moment; is the discretized time interval; is the system frequency variation; for k Total system inertia requirement at any moment; for k The power disturbance the system experiences at that moment; for k The power response of the system at that moment, that is, the power adjustment amount made by the system in response to the power disturbance; for k Virtual power response at each moment; for k The power response quantity related to the unit inertia at any time.

2. The hybrid optimization method for island-type heterogeneous source-load-storage inertia assessment according to claim 1 is characterized in that: The spatiotemporal distribution discrete model is modified in real time based on the difference between the unit frequency and the inertia center frequency according to the inertia distribution coefficient. If the difference between the unit frequency and the inertia center frequency is large, it indicates that the unit inertia is unevenly distributed, and the total system inertia demand and the power response related to the unit inertia need to be adjusted. If the distribution coefficient approaches 1, it indicates that the unit inertia is evenly distributed and no adjustment is required. The specific formula of the inertia distribution coefficient is: ； in, is the system frequency based on the center of inertia; for k moment Inertia distribution coefficient; N is the total number of units in the system; For the i Units in k The frequency of the moment.

3. The hybrid optimization method for island-type heterogeneous source-load-storage inertia assessment according to claim 2 is characterized in that: described k The power response quantity related to the unit inertia at the moment is obtained by k The load power at each moment is corrected and the specific formula is: ； in, is the load voltage equivalent inertia correction; is the reference load power; for k The voltage value of the system at the moment, is the reference voltage value; is constant impedance; is the constant current load ratio.

4. The hybrid optimization method for island-type heterogeneous source-load-storage inertia assessment according to claim 1 is characterized in that: The first constraint in S3 is the RoCoF constraint, and its specific formula is: ； in, is the value of the system frequency change rate based on the inertia center at time k, reflecting how fast the frequency changes with time; The maximum frequency change rate threshold allowed is used to limit the frequency change speed and ensure system stability.

5. The hybrid optimization method for island-type heterogeneous source-load-storage inertia assessment according to claim 1 is characterized in that: The specific process of further processing the solution set using the genetic algorithm in S4 is: using the penalty function to process the constraints and then solve the optimal solution. The specific formula is: ； in, For k The penalty function value at the moment is used to measure the degree to which the solution violates the constraints; α is the penalty coefficient, which is used to adjust the sensitivity of the penalty function to constraint violations; is the rate of change of system frequency based on the center of inertia k The value of the moment; The maximum frequency change rate threshold allowed.

6. The hybrid optimization method for island-type heterogeneous source-load-storage inertia assessment according to claim 1 is characterized in that: The second constraint in S5 includes the frequency minimum point constraint and the energy storage frequency capacity constraint; wherein the frequency minimum point constraint is that the actual value of the system frequency minimum point is greater than or equal to the minimum frequency value allowed by the system; the energy storage frequency capacity constraint is that k The power change of the energy storage system participating in frequency regulation at any moment is less than or equal to the maximum frequency regulation power capacity allowed by the energy storage system. The specific formula is: ； ； in, is the actual value of the lowest frequency point of the system; is the lowest frequency value allowed by the system; For k The power change of the energy storage system participating in frequency regulation at each moment; Adjust the power capacity for the maximum frequency allowed by the energy storage system.

7. The hybrid optimization method for island-type heterogeneous source-load-storage inertia assessment according to claim 6 is characterized in that: The actual value of the lowest frequency point of the system is obtained by the sliding window extreme value detection method. First, the time length of the sliding window is set; then the time of the lowest frequency point is determined based on the time length of the sliding window; finally, the frequency value at that moment is obtained by the sliding window extreme value detection method. Calculate the actual lowest frequency , the specific formula is: ； in, is the moment of lowest frequency; is the system frequency based on the center of inertia; k is the index value of the discrete time series; is the index value corresponding to the starting time of the sliding window; T is the length of the sliding window.

8. An island-type heterogeneous source-load-storage inertia assessment hybrid optimization system, characterized by: The system applies a hybrid optimization method for evaluating inertia of an isolated heterogeneous source-load-storage according to any one of claims 1 to 7, and the system includes a model building module, a solution set generation and preliminary processing module, and a solution set optimization and judgment module; The model building module is used to model multi-source inertia, obtain and build a spatiotemporal distribution discrete model based on the multi-source inertia dynamic model and related parameters; The solution set generation and preliminary processing module is used to generate an initial solution set of the spatiotemporal distribution discrete model using an ant colony algorithm, and determine whether the solution set satisfies the first constraint, and if not, return to regenerate; The solution set optimization and judgment module is used to further process the solution set using a genetic algorithm, output an optimized solution set, and judge whether the optimized solution set meets the second constraint. If so, the optimal inertia configuration is output; if not, the optimal inertia configuration is returned to continue optimization.

Images