Graded protection constant value optimization method for power distribution network containing IIDG and energy storage

Abstract

The invention discloses an IIDG and energy storage-containing power distribution network grading protection constant value optimization method, which comprises the following steps: S1, establishing an IIDG and energy storage access-containing power distribution network grading protection constant value multi-target optimization model, specifically, constructing an optimization model with the minimum total action time, the highest protection sensitivity and the minimum load loss as targets, upper and lower protection cooperation constraints are considered; s2, an improved multi-target particle swarm algorithm is adopted for solving, specifically, chaotic mapping is introduced to enhance population diversity, an adaptive splitting strategy is adopted to improve solution set distribution, and a Pareto optimal solution set covering multiple tradeoff schemes is obtained after solving; s3, using k-means clustering analysis to screen out a global optimal protection constant value scheme from the Pareto optimal solution set; the method can effectively coordinate protection quickness, sensitivity and selectivity, is superior to a traditional optimization algorithm, and provides a feasible scheme for protection setting of the power distribution network under high-proportion new energy access.

Description

Technical Field

[0001] This invention belongs to the field of distribution network protection setting technology, specifically relating to a method for optimizing the hierarchical protection settings of distribution networks including IIDG and energy storage. Background Technology

[0002] The penetration rate of inverter-interfaced distributed generation (IIDG), represented by photovoltaic and wind power, and energy storage systems (ESS) in distribution networks continues to rise, gradually giving rise to a new form in which distribution networks are connected to a high proportion of power electronic devices. IIDG and ESS are connected to the grid through power electronic converters, and their operating characteristics have profoundly changed the traditional unidirectional radial power flow pattern and fault characteristics of distribution networks, bringing unprecedented challenges to the safe and stable operation of the system.

[0003] From the perspective of distribution network protection and control, the integration of IIDG and ESS significantly alters the amplitude, phase, and direction characteristics of the short-circuit current after a fault. IIDG, influenced by Low Voltage Ride-Through (LVRT) control strategies and output current limiting, provides a limited short-circuit current that exhibits a non-linear relationship with the grid voltage. ESS, due to its source / load duality, displays "load" and "power source" characteristics for the fault current during charging and discharging, respectively, and its fault current is also controlled by the converter strategy. This complex fault characteristic leads to a serious risk of maloperation or failure to operate for traditional overcurrent-based protection devices, especially inverse-time overcurrent protection (ITOC). Furthermore, the "boosting" or "draining" effects of IIDG and ESS alter the current flowing through the protection device, disrupting the timing coordination between upstream and downstream protection systems and weakening the selectivity, speed, and sensitivity of the protection.

[0004] Although traditional three-stage current protection has advantages such as simple structure and convenient setting in radial distribution networks, its setting principle is based on short-circuit current calculations under the system's maximum / minimum operating conditions, and does not fully consider the spatiotemporal variability and control dependence of fault currents after the integration of IIDG and ESS. Existing research shows that in distribution networks with a high proportion of IIDG and ESS integration, the range of traditional current protection may extend to adjacent lines, leading to overlapping protection ranges and triggering cascading trips; or the protection range may be severely reduced under certain operating conditions, resulting in protection failure. In addition, the dynamic response of IIDG and ESS during faults (such as current support during LVRT and converter disconnection processes) further increases the complexity of protection coordination.

[0005] To address these challenges, scholars have proposed various methods for optimizing protection settings. For example, Wang Gang et al. established a multi-objective optimization model for inverse-time overcurrent protection considering speed and sensitivity, and used an improved multi-objective particle swarm optimization (MOPSO) algorithm for setting tuning. Luan Kun et al. constructed a constrained multi-objective optimization model with the objectives of minimizing overall fault clearing time, protection sensitivity, and load loss, and combined it with a clustering algorithm to select the final setting scheme from the Pareto solution set. These studies have improved the coordination performance of protection to some extent, but still have problems such as insufficient initial population diversity, uneven distribution of solution set, and susceptibility to local optima. Furthermore, the model construction did not fully consider the impact of various operating states of IIDG and ESS (such as ESS charging and discharging switching and IIDG disconnection process) on the dynamic adaptability of protection settings.

[0006] To address the aforementioned issues, it is essential to develop a method for optimizing the hierarchical protection settings of distribution networks that incorporates IIDG and energy storage. Summary of the Invention

[0007] The purpose of this invention is to overcome the shortcomings of the prior art and provide an offline setting and online calling method for optimizing the protection settings of distribution networks containing IIDG and energy storage. This method can effectively coordinate the trade-off between protection speed, sensitivity and selectivity, and can significantly improve the adaptability and coordination of protection in active distribution networks under different operating scenarios. It also provides an effective way to solve the protection selectivity problem caused by the access of new energy sources.

[0008] The objective of this invention is achieved as follows: a method for optimizing the hierarchical protection settings of a distribution network including IIDG and energy storage, comprising the following steps:

[0009] S1. Establish a multi-objective optimization model for the hierarchical protection settings of the distribution network including IIDG and energy storage access:

[0010] S11, Determine the optimization variables: Take the current setting value Iset and time setting value T of each protection segment in the system as the variables to be optimized;

[0011] S12, Construct the objective function: minimize the total operating time of all protections under various faults, maximize the minimum sensitivity coefficient of all protections under various faults, and minimize the total load disconnected due to protection operation;

[0012] S13, Set constraints: including coordination constraints between upper and lower protection sections of the same protection, selective coordination constraints between different protections, and physical upper and lower limit constraints of setting values;

[0013] S2 is solved using an improved multi-objective particle swarm optimization algorithm: the standard multi-objective particle swarm optimization algorithm is improved by using chaotic initialization and adaptive splitting strategies, and a Pareto optimal solution set covering multiple trade-off schemes is obtained after solving the algorithm.

[0014] S3. Determine the final protection setting scheme based on cluster analysis: First, in the three-dimensional target space, the Pareto optimal solution set is clustered into speed priority class, balance class and selectivity priority class respectively. Then, the selectivity priority class is selected as the highest priority and the sensitivity is selected as the second priority for screening. Finally, the solution with the shortest total action time is selected from the selected solutions as the final global optimal protection setting scheme.

[0015] Preferably, in step S11, the current protection achieves mutual coordination between adjacent protections at different levels through two fixed values: current and operating time, thereby ensuring the selectivity of the operation. The instantaneous overcurrent protection does not have an operating delay, so the current protection I time period is not considered when optimizing the variables. Let the variables be as shown in equation (1):

[0016] (1);

[0017] In the formula, d is the protection code, i.e. The protection setting for stage I is used to protect d; m is the total number of protections configured in the distribution network; stage I and stage III protections are configured on the penultimate level of the power grid.

[0018] Preferably, in step S12, three objective functions are constructed based on the four characteristics of relay protection: "reliability," "selectivity," "speed," and "sensitivity," as follows:

[0019] S121, Operating Time Limit: The total operating time of the protection system during M short circuits in the distribution network is used as the speed performance indicator. The objective function is established as follows:

[0020] (2);

[0021] In the formula, M represents the operating time of each level of protection, and M represents the number of short-circuit faults. This represents the total time of all protection actions during the a-th short circuit.

[0022] S122, Sensitivity: The minimum sensitivity coefficient of the protection in the system is taken as the optimization target, that is:

[0023] (3);

[0024] In the formula, This represents the short-circuit current during the a-th fault. This is the current protection setting value for the b-th protection segment c that takes protective action during the a-th fault.

[0025] S123, Reliability and Selectivity: When a fault occurs in the power system, the protection should minimize load loss, that is, disconnect the fault from the power system in the shortest interval. Based on this, the evaluation index is as shown in Equation (4):

[0026] (4);

[0027] In the formula, This indicates whether the b-th protection device activates during the a-th short circuit. For action, Inaction; The load that is disconnected by the protection system during the a-th short circuit for normal operation; The load for the protection of the a-th resection of b; When the system fails, there is no protection action, meaning the entire system load is lost;

[0028] Therefore, the objective function for this type of problem is:

[0029] (5).

[0030] Preferably, in step S13, the constraints are set as follows:

[0031] S131, Constraints on Coordination of Optimized Variables for the Same Protection: The current protection settings for stages I, II, and III of the same protection should decrease progressively; similarly, the operating time of stage III protection should be longer than that of stage II, i.e.:

[0032] (6);

[0033] S132, Overall optimization variable constraint: The line current protection setting value and operating time limit shall not exceed the maximum limit, that is:

[0034] (7);

[0035] In the formula, , These are the maximum current limit and the maximum operating time limit for the b-th protection, respectively. The maximum short-circuit current at the substation outlet is calculated based on the system configuration.

[0036] S133, Constraints on Current Setting Values ​​for Different Protections: To meet the selectivity requirements of relay protection, different protections must coordinate with each other. The upper-level current protection stage II must coordinate with the lower-level current protection stage I or II, and the upper-level current protection stage III must coordinate with the lower-level current protection stage III. The closer the current protection stage is to the power source, the larger its setting value should be.

[0037] (8);

[0038] In the formula, The current protection setting for the upper-level line current stage II is set. Set the protection value for the downstream line current in either stage I or stage II;

[0039] S134, Constraints on the Operating Time of Different Protections: Similarly, the closer to the power source, the longer the operating time of the current protection in the same segment, that is:

[0040] (9);

[0041] In the formula, Set the protection time value for the second stage of the upstream line current. Set the protection time value for the downstream line current stage I or II; considering the protection speed index, take... .

[0042] Preferably, step S2 includes the following steps:

[0043] S21, Improved Chaotic Initialization:

[0044] S211, Generating a chaotic Tent sequence: The mathematical expression for the Tent mapping is shown in equation (10). Considering the chaotic characteristics, its control parameters... Get 2.0, input value The output sequence has uniform traversal. Properties of intervals:

[0045] (10);

[0046] In the formula, Let be the chaotic value in the m-th iteration; The chaotic value in the (m+1)th iteration; for two types of optimization variables , Two independent N-dimensional Tent sequences are generated as shown in equation (11), where N is the population size. This represents the m-th iteration chaotic value of the current setting of the j-th segment of the b-th protection. This represents the m-th iteration chaotic value of the b-th protection setting value during the j-th time period;

[0047] (11);

[0048] S212, Mapping the chaotic sequence to the feasible region: As shown in equation (12), the Tent sequence is mapped to the actual search space of the optimization variables;

[0049] (12);

[0050] In the formula, , The current setpoint and action time limit are given for the m-th particle;

[0051] After mapping, it is necessary to further verify whether the particles satisfy the constraints. Infeasible solutions are corrected by chaotic perturbation to ensure the feasibility of the initial population.

[0052] S22, Adaptive Splitting Strategy Design:

[0053] S221, Solution density evaluation: Dynamically replenish sparse region particles based on the density of archived solutions; splitting criterion is:

[0054] (13);

[0055] In the formula, , For the density variance in the archived solution, Density index;

[0056] For non-dominated solutions in the external archive, calculate the k-nearest neighbor crowding distance for each solution. To balance computational load and accuracy, The density index is shown in equation (14):

[0057] (14);

[0058] In the formula, to avoid the denominator being 0, take... ;

[0059] S222, Sparse Region Splitting: For a region that satisfies the splitting condition, select the n solutions with the lowest density as splitting centers, where n is one-fifth of the archive size. For each center solution... This generates two new particles, namely:

[0060] (15);

[0061] In the formula, , .

[0062] Preferably, in step S3, the k-means clustering algorithm is introduced to perform cluster analysis on the Pareto solution set, and the optimal solution scheme is determined based on the cluster centers. The specific algorithm flow for finding the optimal solution is as follows:

[0063] After starting the algorithm process, set various parameters for algorithm operation, including particle swarm size, maximum number of iterations, particle velocity range, learning factor, inertia weight, and density threshold.

[0064] Next, the Tent chaotic map is used to generate the initial particle swarm. The Tent map generates a chaotic sequence through iteration, and then maps the chaotic sequence to the feasible solution domain of the problem to obtain the initial particle positions, ensuring that the initial population is uniform and comprehensive within the feasible domain.

[0065] The next step is to check the constraints of each particle in the initial population to ensure that its position meets the constraints of the problem, and to correct the position of particles that do not meet the constraints.

[0066] The next step is to evaluate the fitness of the corrected particles by calculating the fitness value of each particle under the objective function in order to assess the quality of the solution.

[0067] Next, an external archive is created and initialized to store the currently found non-dominated solutions, i.e., Pareto optimal solutions; after initialization, the main iteration loop begins.

[0068] In the main iteration loop, the iteration count is checked first to determine if the current iteration count has reached the maximum value. If it has reached the maximum value, the loop is exited and the Pareto optimal solution set output step is executed; otherwise, the iteration loop step is continued.

[0069] In the next iteration, the historical best solution for each particle is updated based on the current fitness value, while a global best solution is selected from the external archive.

[0070] In the next step of the iterative loop, the updated particle position is constrained and corrected if necessary to ensure that the particle is always within the feasible region.

[0071] The next step in the iterative loop is to evaluate the fitness of the corrected particles, calculate the fitness value of each particle under the objective function, and evaluate the quality of the solution.

[0072] In the next step of the iterative loop, the new particle is compared with the solution in the external archive, non-dominated sorting and crowding calculation are performed, and the archive is updated to retain the current non-dominated solution; if the archive size exceeds the limit, it is pruned according to the solution density, and solutions in dense regions are removed.

[0073] The next step in the iterative loop is to calculate the density of solutions in the external archive. The density value reflects the sparsity of the solution distribution in the target space.

[0074] In the next step of the iterative loop, a density condition judgment is performed to check whether the density variance of the archived solution is greater than a preset threshold. If the condition is met, the following sub-step is executed; if the condition is not met, the third sub-step is executed directly.

[0075] The first sub-step is to select the solution with the lowest density from the archive, that is, the solution in the sparsest region.

[0076] The second sub-step involves generating new particles centered on the selected solution using a specific strategy and adding them to the population to enhance the exploration of sparse regions.

[0077] The third sub-step involves comparing the newly generated particles with the solutions in the archive to update the external archive.

[0078] In the next iteration, after completing the sub-step, update the individual particle optimality and the global optimality, and return to determine whether the number of iterations has reached the maximum value;

[0079] Next, after the iteration loop ends, the Pareto optimal solution set stored in the external archive is output as the final result of the algorithm.

[0080] Next, the algorithm terminates.

[0081] Preferably, the method further includes step S4, constructing an adaptive setpoint strategy library, the steps of which are as follows:

[0082] S41, Define typical operating scenarios: Based on the typical output modes of IIDG and energy storage, define a set of representative typical operating scenarios. This scenario set should cover the main characteristics of system operation, including Peak power generation Charging overnight;

[0083] S42, Offline generation of strategy library: For each typical scenario, repeat steps S1, S2 and S3 to obtain the optimal protection setting scheme corresponding to the scenario. The schemes of all typical scenarios constitute the "adaptive protection setting strategy library".

[0084] S43, Online Real-time Adjustment: The system obtains the operating status of the distribution network in real time through the monitoring system, matches the current status with the scenarios in the strategy library, calls the corresponding optimal setpoint, and sends the called new setpoint to each protection device through the communication system, so as to realize the online adaptive adjustment of the protection setpoint.

[0085] More preferably, step S43 includes the following steps:

[0086] S431: During normal operation, the distribution network operation status is obtained, and the remote communication device based on MMS transmits the distribution network operation information to the dispatching terminal.

[0087] S432, calculate IIDG and / or energy storage. The protection setting corresponding to the range, if the fluctuation range of IIDG and / or energy storage output obtained by the dispatching terminal is within... If the output is within the allowed range, no adjustment of the setpoint is required; if the IIDG and / or energy storage output fluctuations obtained by the dispatcher exceed the allowable output fluctuation range, proceed to the next step to update the setpoint.

[0088] S433, Setting update process: The scheduling terminal matches the current running status with typical scenarios in the "Adaptive Protection Setting Strategy Library", calls the corresponding optimal protection setting scheme, and changes the output fluctuation requirements;

[0089] S434: The new setting is transmitted to the protection device through the communication device and compared with the existing setting of the protection device. If the difference exceeds the allowable range, the protection device stores and uses the adjusted setting, realizing the online update of the protection setting.

[0090] Due to the adoption of the above technical solutions, the beneficial effects of this invention are as follows: Based on in-depth analysis of the equivalent models of IIDG and ESS during faults and their influence mechanism on the coordination characteristics of multi-level protection, this invention constructs an optimization model with the objectives of minimizing total operating time, maximizing protection sensitivity, and minimizing load loss, while considering the coordination constraints between upper and lower level protections; simultaneously, to solve the optimization model, an improved multi-objective particle swarm optimization algorithm is adopted, introducing chaotic mapping to enhance population diversity, employing an adaptive splitting strategy to improve the solution set distribution, and using cluster analysis to select the final setting scheme from the Pareto optimal solution; finally, simulation results based on the IEEE 33-bus system show that this invention can effectively coordinate protection speed, sensitivity, and selectivity, outperforming traditional optimization algorithms, and providing a feasible solution for protection setting in distribution networks with a high proportion of new energy access. Attached Figure Description

[0091] Figure 1 This is a flowchart of the method steps of the present invention.

[0092] Figure 2 This is a flowchart of the solution algorithm of the present invention.

[0093] Figure 3 This is a network topology diagram for a single IIDG access during the impact analysis of this invention.

[0094] Figure 4 This is a flowchart of the current iteration calculation for the impact analysis of this invention.

[0095] Figure 5 This is a schematic diagram of the IEEE 33-node system analyzed in the example of this invention.

[0096] Figure 6 This is a diagram showing the Pareto front classification results from the example analysis of this invention.

[0097] Figure 7 This is the Pareto front projection diagram of the XY plane in the numerical analysis of this invention.

[0098] Figure 8 This is the Pareto front plot obtained from the standard MOPSO in the numerical analysis of this invention.

[0099] Figure 9 This is a graph showing the final value optimization results of the case study analysis of this invention. Detailed Implementation

[0100] The technical solution of the present invention will be further described in detail below through embodiments and in conjunction with the accompanying drawings.

[0101] like Figure 1 and Figure 2 As shown, this invention provides a method for optimizing the hierarchical protection settings of a distribution network including IIDG and energy storage, comprising the following steps:

[0102] S1. Establish a multi-objective optimization model for the hierarchical protection settings of the distribution network including IIDG and energy storage access.

[0103] S11, Determine the optimization variables: Take the current setting value Iset and time setting value T of each protection segment in the system as the variables to be optimized;

[0104] Current protection achieves mutual coordination between adjacent protection levels through two set values: current and operating time, thereby ensuring the selectivity of operation. Instantaneous overcurrent protection does not have an operating delay, so the current protection time period is not considered when optimizing variables. Let the variables be as shown in equation (1):

[0105] (1);

[0106] In the formula, d is the protection code, i.e. The protection setting for stage I is used to protect d; m is the total number of protections configured in the distribution network; in the penultimate level of the power grid, generally only stage I and stage III protections can be configured.

[0107] S12, Constructing the objective function: Traditional three-stage current protection is mainly designed for one protection. In actual systems, there are multiple protections. In order to achieve hierarchical protection between multiple line segments, it is necessary to conduct an overall evaluation of the protection performance of the system, including minimizing the total operating time of all protections under various faults, maximizing the minimum sensitivity coefficient of all protections under various faults, and minimizing the total load disconnected due to protection operation.

[0108] This invention constructs three objective functions based on four characteristics of relay protection: reliability, selectivity, speed, and sensitivity, to measure the rationality of the hierarchical protection configuration of distribution networks with IIDG and energy storage access, as detailed below:

[0109] S121, Operating time limit: The rapid operation of the protection can effectively suppress the impact and damage of short-circuit faults on expensive main equipment such as generators and transformers, and ensure asset safety; secondly, it is the primary condition for maintaining the transient stability of the power system, and can effectively prevent the system from losing synchronization and avoid large-scale power outage accidents.

[0110] This invention uses the total protection action time for M short circuits in the distribution network as the speed performance index, and establishes the objective function as follows:

[0111] (2);

[0112] In the formula, M represents the operating time of each level of protection, and M represents the number of short-circuit faults. This represents the total time of each protection action during the a-th short circuit.

[0113] S122, Sensitivity: Regardless of how the system operates or the location and type of the short circuit, the protection should be sensitive to and react correctly when a short circuit occurs.

[0114] This invention takes the minimum sensitivity coefficient protected in the system as the optimization target, that is:

[0115] (3);

[0116] In the formula, This represents the short-circuit current during the a-th fault. This is the current protection setting value for the b-th protection segment c when the protection action is taken during the a-th fault.

[0117] S123, Reliability and Selectivity: Reliability is the most fundamental requirement for relay protection, which includes two aspects: safety and reliability, namely, no false operation and no failure to operate.

[0118] When a fault occurs in the power system, the protection should minimize load loss, that is, disconnect the fault from the power system in the shortest possible interval. Based on this, the evaluation index is as shown in equation (4):

[0119] (4);

[0120] In the formula, This indicates whether the b-th protection device activates during the a-th short circuit. For action, Inaction; The load that is disconnected by the protection system during the a-th short circuit for normal operation; The load for the protection of the a-th resection of b; When the system fails, there is no protection action, meaning the entire system load is lost;

[0121] Therefore, the objective function for this type of problem is:

[0122] (5).

[0123] S13, Set constraints: including coordination constraints between upper and lower protection sections of the same protection, selective coordination constraints between different protections, and physical upper and lower limit constraints of setting values.

[0124] S131, Constraints on Coordination of Optimized Variables for the Same Protection: The current protection settings for stages I, II, and III of the same protection should decrease progressively; similarly, the operating time of stage III protection should be longer than that of stage II, i.e.:

[0125] (6).

[0126] S132, Overall optimization variable constraint: The line current protection setting value and operating time limit shall not exceed the maximum limit, that is:

[0127] (7);

[0128] In the formula, , These are the maximum current limit and the maximum operating time limit for the b-th protection, respectively. The maximum short-circuit current at the substation outlet is calculated based on the system configuration.

[0129] S133, Constraints on Current Setting Values ​​for Different Protections: To meet the selectivity requirements of relay protection, different protections must coordinate with each other. The upper-level current protection stage II must coordinate with the lower-level current protection stage I or II, and the upper-level current protection stage III must coordinate with the lower-level current protection stage III. The closer the current protection stage is to the power source, the larger its setting value should be.

[0130] (8);

[0131] In the formula, The current protection setting for the upper-level line current stage II is set. The current protection setting for the downstream line current, either Section I or Section II.

[0132] S134, Constraints on the Operating Time of Different Protections: Similarly, the closer to the power source, the longer the operating time of the current protection in the same segment, that is:

[0133] (9);

[0134] In the formula, Set the protection time value for the second stage of the upstream line current. Set the protection time value for the downstream line current stage I or II; considering the protection speed index, take... .

[0135] S2 is solved using an improved multi-objective particle swarm optimization algorithm: the standard multi-objective particle swarm optimization algorithm is improved by using chaotic initialization and adaptive splitting strategies, and a Pareto optimal solution set covering multiple trade-off schemes is obtained after solving the algorithm.

[0136] Traditional multi-objective particle swarm optimization (MOPSO) algorithms suffer from insufficient initial population diversity, uneven distribution of external archived solutions, and susceptibility to local optima. Therefore, this invention proposes a chaotic-adaptive splitting multi-objective particle swarm optimization algorithm. It enhances the exploratory nature of the population with Tent chaotic mapping and optimizes the distribution of non-dominated solutions with an adaptive splitting strategy, ultimately achieving multi-objective optimization of the protection setpoint with "minimum action time limit, highest sensitivity, and minimum load loss".

[0137] S21, Improved Chaotic Initialization: Traditional MOPSO uses random initialization, which easily leads to the initial particles gathering in local regions. Especially in the optimization of protected setpoints, since the setpoints need to meet multiple constraints such as upper and lower level cooperation and IIDG output constraints, random initialization is prone to generating a large number of infeasible solutions. Therefore, this invention introduces Tent chaotic mapping to generate the initial population, and uses its ergodicity and randomness to improve the population diversity.

[0138] S211, Generating a chaotic Tent sequence: The mathematical expression for the Tent mapping is shown in equation (10). Considering the chaotic characteristics, its control parameters... Get 2.0, input value The output sequence has uniform traversal. Properties of intervals:

[0139] (10);

[0140] In the formula, Let be the chaotic value in the m-th iteration; The chaotic value in the (m+1)th iteration; for two types of optimization variables , Two independent N-dimensional Tent sequences are generated as shown in equation (11), where N is the population size. This represents the m-th iteration chaotic value of the current setting of the j-th segment of the b-th protection. This represents the m-th iteration chaotic value of the b-th protection setting value during the j-th time period;

[0141] (11).

[0142] S212, Mapping the chaotic sequence to the feasible region: As shown in equation (12), the Tent sequence is mapped to the actual search space of the optimization variables;

[0143] (12);

[0144] In the formula, , The current setpoint and action time limit are given for the m-th particle;

[0145] After mapping, it is necessary to further verify whether the particles satisfy the constraints. Infeasible solutions are corrected by chaotic perturbation to ensure the feasibility of the initial population.

[0146] S22, Adaptive Splitting Strategy Design:

[0147] S221, Solution Density Assessment: External archiving is the core of MOPSO's storage of non-dominated solutions. Traditional methods only truncate the archive based on congestion distance, which easily leads to insufficient solution distribution in sparse regions (such as the protection setpoint solution in IIDG high-output scenarios). Therefore, this invention dynamically supplements sparse region particles based on the density of archived solutions. The splitting criterion is:

[0148] (13);

[0149] In the formula, , For the density variance in the archived solution, This is a density index.

[0150] For non-dominated solutions in the external archive, calculate the k-nearest neighbor crowding distance for each solution. To balance computational load and accuracy, The density index is shown in equation (14):

[0151] (14);

[0152] In the formula, to avoid the denominator being 0, take... .

[0153] S222, Sparse Region Splitting: For a region that satisfies the splitting condition, select the n solutions with the lowest density as splitting centers, where n is one-fifth of the archive size. For each center solution... This generates two new particles, namely:

[0154] (15);

[0155] In the formula, , .

[0156] S3. Determine the final protection setting scheme based on cluster analysis: First, in the three-dimensional target space, the Pareto optimal solution set is clustered into speed priority class, balance class and selectivity priority class respectively. Then, the selectivity priority class is selected as the highest priority and the sensitivity is selected as the second priority for screening. Finally, the solution with the shortest total action time is selected from the selected solutions as the final global optimal protection setting scheme.

[0157] Multi-objective optimization algorithms ultimately generate a Pareto optimal solution set, which contains many feasible solutions that are not dominated by any of the three objectives. To select a final protection setting scheme with the best overall performance and ease of engineering application from this solution set, this invention introduces the k-means clustering algorithm to perform cluster analysis on the Pareto solution set and determines the optimal solution scheme based on the cluster centers. The specific algorithm flow for finding the optimal solution is as follows:

[0158] After starting the algorithm process, set various parameters for algorithm operation, including particle swarm size, maximum number of iterations, particle velocity range, learning factor, inertia weight, and density threshold.

[0159] Next, the Tent chaotic map is used to generate the initial particle swarm. The Tent map generates a chaotic sequence through iteration, and then maps the chaotic sequence to the feasible solution domain of the problem to obtain the initial particle positions, ensuring that the initial population is uniform and comprehensive within the feasible domain.

[0160] The next step is to check the constraints of each particle in the initial population to ensure that its position meets the constraints of the problem, and to correct the position of particles that do not meet the constraints.

[0161] The next step is to evaluate the fitness of the corrected particles by calculating the fitness value of each particle under the objective function in order to assess the quality of the solution.

[0162] Next, an external archive is created and initialized to store the currently found non-dominated solutions, i.e., Pareto optimal solutions; after initialization, the main iteration loop begins.

[0163] In the main iteration loop, the iteration count is checked first to determine if the current iteration count has reached the maximum value. If it has reached the maximum value, the loop is exited and the Pareto optimal solution set output step is executed; otherwise, the iteration loop step is continued.

[0164] In the next iteration, the historical best solution for each particle is updated based on the current fitness value, while a global best solution is selected from the external archive.

[0165] In the next step of the iterative loop, the updated particle position is constrained and corrected if necessary to ensure that the particle is always within the feasible region.

[0166] The next step in the iterative loop is to evaluate the fitness of the corrected particles, calculate the fitness value of each particle under the objective function, and evaluate the quality of the solution.

[0167] In the next step of the iterative loop, the new particle is compared with the solution in the external archive, non-dominated sorting and crowding calculation are performed, and the archive is updated to retain the current non-dominated solution; if the archive size exceeds the limit, it is pruned according to the solution density, and solutions in dense regions are removed.

[0168] The next step in the iterative loop is to calculate the density of solutions in the external archive. The density value reflects the sparsity of the solution distribution in the target space.

[0169] In the next step of the iterative loop, a density condition judgment is performed to check whether the density variance of the archived solution is greater than a preset threshold. If the condition is met, the following sub-step is executed; if the condition is not met, the third sub-step is executed directly.

[0170] The first sub-step is to select the solution with the lowest density from the archive, that is, the solution in the sparsest region.

[0171] The second sub-step involves generating new particles centered on the selected solution using a specific strategy and adding them to the population to enhance the exploration of sparse regions.

[0172] The third sub-step involves comparing the newly generated particles with the solutions in the archive to update the external archive.

[0173] In the next iteration, after completing the sub-step, update the individual particle optimality and the global optimality, and return to determine whether the number of iterations has reached the maximum value;

[0174] Next, after the iteration loop ends, the Pareto optimal solution set stored in the external archive is output as the final result of the algorithm.

[0175] Next, the algorithm terminates.

[0176] S4, Construct an adaptive setpoint strategy library:

[0177] The aforementioned optimization model aims to determine the optimal solution for protection settings for a specific operating mode. However, the operating state of distribution networks with a high proportion of IIDG and energy storage has significant time-varying characteristics, and a single fixed setting scheme is difficult to maintain optimal performance in all weather conditions and scenarios. To solve this problem, this paper proposes a method for constructing an adaptive protection setting strategy library based on multiple operating scenarios.

[0178] S41, Define typical operating scenarios: Based on the typical output modes of IIDG and energy storage, define a set of representative typical operating scenarios. This scenario set should cover the main characteristics of system operation, such as:

[0179] (Peak power generation): Rated output of photovoltaic and wind turbines, maximum discharge state of energy storage;

[0180] (Nighttime charging): Photovoltaic output is zero, wind turbine output is at its lowest level, and energy storage is at its maximum charging state.

[0181] S42, Offline generation of strategy library: For each typical scenario, repeat steps S1, S2 and S3 to obtain the optimal protection setting scheme corresponding to the scenario. The schemes of all typical scenarios constitute the "adaptive protection setting strategy library".

[0182] S43, Online Real-Time Adjustment: The system acquires the real-time operating status of the distribution network through a monitoring system, matches the current status with scenarios in the strategy library, calls the corresponding optimal setpoint, and distributes the new setpoint to each protection device via the communication system, achieving online adaptive adjustment of protection setpoints. The specific process is as follows:

[0183] S431: During normal operation, the distribution network operation status is obtained, and the remote communication device based on MMS transmits the distribution network operation information to the dispatching terminal.

[0184] S432, calculate IIDG and / or energy storage. The protection setting value corresponding to the range (taking into account a certain output fluctuation) is if the IIDG and / or energy storage output fluctuation range obtained by the dispatching terminal is within... If the output is within the allowed range, no adjustment of the setpoint is required; if the IIDG and / or energy storage output fluctuations obtained by the dispatcher exceed the allowable output fluctuation range, proceed to the next step to update the setpoint.

[0185] S433, Setting update process: The scheduling terminal matches the current running status with typical scenarios in the "Adaptive Protection Setting Strategy Library", calls the corresponding optimal protection setting scheme, and changes the output fluctuation requirements;

[0186] S434: The new setting is transmitted to the protection device through the communication device and compared with the existing setting of the protection device. If the difference exceeds the allowable range, the protection device stores and uses the adjusted setting, realizing the online update of the protection setting.

[0187] The following is an analysis of the impact of distributed generation access on distribution network protection, taking IIDG as an example.

[0188] I. Short-circuit fault output characteristics of distributed power sources:

[0189] After a fault, the distributed power source adopts a constant power control strategy with low voltage ride-through capability. When a fault occurs in the distribution network, it can prioritize outputting reactive power and provide a certain voltage as support for the grid connection point.

[0190] To prevent damage to the power electronic equipment in the grid-connected system, the amplitude of the IIDG output current should be limited to within 1.5 times the rated current. Therefore, the IIDG can be equivalent to a positive sequence current source controlled by the grid connection point voltage. During a fault, its output reactive current is shown in equation (16):

[0191] (16);

[0192] In the formula, , , Rated current and voltage, Output reactive current to IIDG The voltage at the grid connection point. This is a reference value for active power.

[0193] After a fault occurs in the IIDG, within the current limiting range, its output active current is as shown in equation (17):

[0194] (17);

[0195] In the formula, Output reactive current to IIDG This represents the positive sequence voltage of the grid connection point d-axis during a fault.

[0196] The amplitude and lag phase of the fault current emitted by the IIDG are as follows:

[0197] (18);

[0198] Distributed energy storage systems absorb active power while simultaneously injecting reactive power into the system during charging; conversely, during discharging, the active power is reversed compared to charging, injecting both active and reactive power into the system simultaneously. Furthermore, energy storage power stations employ control strategies to suppress negative-sequence components, ensuring that only positive-sequence components exist in their fault currents, regardless of whether the energy storage was in a discharging or charging state before the fault. According to the national standard GB / T 34120-2017 "Technical Specification for Energy Storage Converters in Electrochemical Energy Storage Systems," energy storage can prioritize outputting reactive power and providing a certain voltage support to the grid connection point during distribution network faults. Therefore, the current output characteristics of distributed energy storage systems during faults are similar to those of IIDGs (Integrated Induction Generators).

[0199] II. The impact of distributed power source access location on distribution network line protection:

[0200] 1. Impact of a single IIDG connection on different fault locations:

[0201] like Figure 3The diagram shows a distribution network model with IIDG access, in which two faults f1 and f2 are set.

[0202] Downstream faults of IIDG have a significant impact on protection. When a three-phase short circuit occurs at the end of line L3, the boosting effect of IIDG will extend the downstream current protection range of its connection point, which may lead to failure to operate. The sensitivity of the line protection upstream of its connection point will decrease due to the external drain effect. However, since the setting value of the III-stage protection is relatively low, the connection of IIDG generally will not cause it to fail to operate. When a two-phase short circuit occurs at the end of line L3, IIDG will extend the upstream and downstream current protection range of its connection point. However, since the fault current of a two-phase short circuit is relatively low, the connection of IIDG generally will not cause it to maloperate.

[0203] When a three-phase short-circuit fault occurs at point f2 on line L1, the fault current flowing through protection 1 is provided solely by the main power grid side, and the fault current flowing through protection 4 is provided solely by the IIDG side. In this case, the connection of the IIDG will not affect protection 1. However, because the amplitude of the short-circuit current provided by the IIDG is too small, protection 4 will find it difficult to determine whether the fault is inside or outside the protection zone based on the current magnitude. If a two-phase short-circuit fault occurs at point f2, the connection of the IIDG generally will not cause it to malfunction. However, because the amplitude of the short-circuit current provided by the IIDG to protection P4 is too small, protection P4 will find it difficult to determine whether the fault is inside or outside the protection zone based on the current magnitude.

[0204] 2. Impact of different fault locations when multiple IIDGs are connected:

[0205] In scenarios with multiple IIDGs, the specific impact is determined by the magnitude and location of the fault current output by the IIDG. The fault current may exhibit complex characteristics of multiple sources superimposing or canceling each other out, further increasing the difficulty of judging the fault using the current amplitude. This will not be analyzed in detail here.

[0206] III. Iterative Calculation of Fault Current in Distribution Networks with Distributed Power Generation Access:

[0207] To adjust the current protection settings for the downstream lines of the IIDG, the lines between IIDGs, and adjacent feeders, it is necessary to calculate the fault current flowing through each protection device when a three-phase short circuit occurs at the end of the corresponding line. However, the control strategy of new energy power sources is complex, and there is a strong nonlinear coupling relationship between its output current and the grid connection point voltage. Furthermore, converting complex distribution network lines into equivalent circuit diagrams is difficult. Therefore, iterative methods are needed to solve for the short-circuit current. This invention uses a quasi-Newton iterative algorithm to construct a matrix recursive relationship to replace the Jacobian matrix, maintaining a faster convergence speed while avoiding the complex differentiation and inversion operations in Newton's method. The IIDG employs a low-voltage ride-through and negative-sequence current suppression control method.

[0208] 3.1 Establishment of the nonlinear equation system:

[0209] According to KCL theorem, short-circuit calculation can be transformed into solving a system of nonlinear equations at the nodes:

[0210] (19);

[0211] In the formula, ; It is a function of voltage and includes the nonlinear response of new energy sources; The network node admittance matrix; This represents the node voltage vector.

[0212] This invention uses the inverse Broyden rank-1 method to update The quasi-Newton method has the following iterative format:

[0213] (20);

[0214] In the formula, k represents the number of iterations. Let be the quasi-Newton matrix for the k-th iteration.

[0215] 3.2 Algorithm Flow:

[0216] First, initialize the system admittance matrix and initial voltage value, and enter the iterative loop. The initial voltage value is the rated voltage of the system. Then, calculate the voltage correction using the quasi-Newton method according to equation (17) and update the system state. Iterate until the node voltage difference meets the convergence accuracy. or The final output includes the fault current and the total electrical quantity of the network. The specific process is as follows: Figure 4 As shown.

[0217] The following is a computational example analysis provided by this invention.

[0218] I. Initial parameter settings:

[0219] To comprehensively simulate actual fault scenarios in the power distribution network and construct a representative fault sample set, the present invention employs the following design in the Matlab simulation platform: First, fault points are selected based on the principle of uniform distribution to ensure coverage of key locations such as the beginning and end of the line and the protection zone. To improve the balance between computational efficiency and accuracy, short-circuit faults are deployed at one-third and two-thirds of the line length between two adjacent nodes, respectively. Second, in terms of fault type settings, the ratio of two-phase short circuits to three-phase short circuits is configured as 3:2.

[0220] Adopting such Figure 5The IEEE 33-node system shown was simulated to verify the feasibility and superiority of the model and algorithm. The distribution network is configured with six protection systems: protection systems 1, 2, and 3 are main line protections, and protection systems 4, 5, and 6 are branch protections. A 600kW photovoltaic system is connected at nodes 7 and 20, and a 400kW wind turbine is connected at nodes 16 and 24. Other parameter settings are shown in Table 1.

[0221] .

[0222] II. Protection and Optimization Results:

[0223] To visually demonstrate the solution effect of the multi-objective optimization algorithm and to deeply analyze the inherent trade-offs among the three optimization objectives, the initial number of clusters was selected as 3, as follows: Figure 6 The diagram shows the distribution of the final Pareto front in the three-dimensional target space.

[0224] The Pareto optimal solution integrated surface is distributed in a three-dimensional space spanned by the three objective function axes, rather than clustered in a local region. This distribution shape intuitively confirms that there is a significant competition and trade-off relationship between the three optimization objectives of total action time, minimum sensitivity coefficient and load loss. No single solution can achieve optimality on all objectives simultaneously. In this case, the point closest to the cluster center is selected as the optimal solution for that class.

[0225] III. Protection Optimization Analysis:

[0226] 1. Reasonableness of the objective function:

[0227] To verify the rationality of the model and algorithm, for example... Figure 7 The Pareto front shown is analyzed. Because the numerical ranges of the various optimization indices differ significantly, it is difficult to discern the local characteristics of the three-dimensional front by direct observation. Therefore, the indices are segmented, and the front is plotted at specific points. The two-dimensional projection diagram within the interval is as follows: Figure 7 As shown. By Figure 7 It can be seen that when When fixed within a certain range, the sensitivity index and the total operating time show a clear nonlinear inverse relationship; that is, the improvement of speed is often accompanied by the decrease of sensitivity. This trend is basically consistent with the inherent "speed-sensitivity" trade-off characteristic in relay protection systems.

[0228] 2. Algorithm effectiveness and excellence:

[0229] To verify the effectiveness and superiority of the improved multi-objective particle swarm optimization (MOPSO) algorithm, it was used together with the standard MOPSO algorithm to solve the multi-objective optimization model described by equations (2) to (9). The Pareto front obtained by this algorithm ( Distribution as follows Figure 8 As shown; the results indicate that, compared to standard MOPSO, the improved MOPSO achieves [the desired effect]. The solution set is more evenly distributed.

[0230] Comparing the three cluster centers obtained by the two algorithms in Table 2, it can be seen that: the improved MOPSO can break out of the search range and find a solution with a smaller objective value; compared with the standard MOPSO, the improved MOPSO makes the objective function... and Significantly reduced, With comparable values, it can effectively solve the problem that standard MOPSO is prone to getting trapped in local optima.

[0231] .

[0232] To further evaluate the algorithm's performance, the standard test function DTLZ1 was selected, with a population size of 100, a maximum number of iterations of 50, and 30 independent runs. The mean values ​​of Inverse Generation Distance (IGD), Generation Distance (GD), Spacing, and running time were used to comprehensively evaluate the algorithm's performance in terms of convergence, distribution, and computational efficiency. As shown in Table 3, compared with traditional methods, the improved MOPSO has the smallest mean values ​​for IGD and Spacing, the shortest running time, and a comparable mean value for GD, indicating that it has better convergence accuracy, uniformity of solution set distribution, and computational efficiency.

[0233] .

[0234] IV. Final Protection Scheme Selection:

[0235] Among the three optimized values ​​obtained by the clustering algorithm, class A has the best speed, class B has moderate performance across all categories, and class C has the best selectivity. This invention sets selectivity as the highest priority; therefore, after obtaining the optimal solutions for the three classes using the K-means algorithm, optimization is performed within the class C range, with sensitivity as the second priority. Finally, as shown... Figure 9 As shown, the solution with the minimum total action time is selected as the final protection scheme.

[0236] In summary, this invention addresses the mismatch problem of traditional current protection in distribution networks caused by high-proportion IIDG and energy storage integration by proposing an improved hierarchical protection setting optimization method that involves "offline setting and online invocation." This method constructs a multi-objective optimization model centered on total operating time, protection sensitivity, and load loss, and employs an improved multi-objective particle swarm optimization algorithm for solving the problem. This effectively coordinates the trade-offs between protection speed, sensitivity, and selectivity. Simulation results demonstrate that, compared to traditional setting methods, this invention significantly improves the adaptability and coordination of protection in active distribution networks under different operating scenarios, providing an effective approach to solving the protection selectivity problem brought about by new energy integration.

[0237] 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 optimizing the hierarchical protection settings of a distribution network including IIDG and energy storage, characterized in that, Includes the following steps: S1. Establish a multi-objective optimization model for the hierarchical protection settings of the distribution network including IIDG and energy storage access: S11, Determine the optimization variables: Take the current setting value Iset and time setting value T of each protection segment in the system as the variables to be optimized; S12, Construct the objective function: minimize the total operating time of all protections under various faults, maximize the minimum sensitivity coefficient of all protections under various faults, and minimize the total load disconnected due to protection operation; S13, Set constraints: including coordination constraints between upper and lower protection sections of the same protection, selective coordination constraints between different protections, and physical upper and lower limit constraints of setting values; S2 is solved using an improved multi-objective particle swarm optimization algorithm: the standard multi-objective particle swarm optimization algorithm is improved by using chaotic initialization and adaptive splitting strategies, and a Pareto optimal solution set covering multiple trade-off schemes is obtained after solving the algorithm. S3. Determine the final protection setting scheme based on cluster analysis: First, in the three-dimensional target space, the Pareto optimal solution set is clustered into speed priority class, balance class and selectivity priority class respectively. Then, the selectivity priority class is selected as the highest priority and the sensitivity is selected as the second priority for screening. Finally, the solution with the shortest total action time is selected from the selected solutions as the final global optimal protection setting scheme.

2. The method for optimizing the hierarchical protection settings of a distribution network including IIDG and energy storage according to claim 1, characterized in that, In step S11, the current protection achieves mutual coordination between adjacent protections at different levels through two fixed values: current and operating time, thereby ensuring the selectivity of the operation. The instantaneous overcurrent protection does not have an operating delay, so the current protection I time period is not considered when optimizing the variables. Let the variables be as shown in equation (1): （1）； In the formula, d is the protection code, i.e. The protection setting for stage I is used to protect d; m is the total number of protections configured in the distribution network; stage I and stage III protections are configured on the penultimate level of the power grid.

3. The method for optimizing the hierarchical protection settings of a distribution network including IIDG and energy storage according to claim 1, characterized in that, In step S12, based on the four characteristics of relay protection—"reliability," "selectivity," "speed," and "sensitivity"—three objective functions are constructed, as follows: S121, Operating Time Limit: The total operating time of the protection system during M short circuits in the distribution network is used as the speed performance indicator. The objective function is established as follows: （2）； In the formula, M represents the operating time of each level of protection, and M represents the number of short-circuit faults. This represents the total time of all protection actions during the a-th short circuit. S122, Sensitivity: The minimum sensitivity coefficient of the protection in the system is taken as the optimization target, that is: （3）； In the formula, This represents the short-circuit current during the a-th fault. This is the current protection setting value for the b-th protection segment c that takes protective action during the a-th fault. S123, Reliability and Selectivity: When a fault occurs in the power system, the protection should minimize load loss, that is, disconnect the fault from the power system in the shortest interval. Based on this, the evaluation index is as shown in Equation (4): （4）； In the formula, This indicates whether the b-th protection device activates during the a-th short circuit. For action, Inaction; The load that is disconnected by the protection system during the a-th short circuit for normal operation; The load for the protection of the b resection in the ath case; When the system fails, there is no protection action, meaning the entire system load is lost; Therefore, the objective function for this type of problem is: （5）。 4. The method for optimizing the hierarchical protection settings of a distribution network including IIDG and energy storage according to claim 1, characterized in that, In step S13, the following constraints are set: S131, Constraints on Coordination of Optimized Variables for the Same Protection: The current protection settings for stages I, II, and III of the same protection should decrease progressively; similarly, the operating time of stage III protection should be longer than that of stage II, i.e.: （6）； S132, Overall optimization variable constraint: The line current protection setting value and operating time limit shall not exceed the maximum limit, that is: （7）； In the formula, , These are the maximum current limit and the maximum operating time limit for the b-th protection, respectively. The maximum short-circuit current at the substation outlet is calculated based on the system configuration. S133, Constraints on Current Setting Values ​​for Different Protections: To meet the selectivity requirements of relay protection, different protections must coordinate with each other. The upper-level current protection stage II must coordinate with the lower-level current protection stage I or II, and the upper-level current protection stage III must coordinate with the lower-level current protection stage III. The closer the current protection stage is to the power source, the larger its setting value should be. （8）； In the formula, The current protection setting for the upper-level line current (Section II) is set. Set the protection value for the downstream line current in either stage I or stage II; S134, Constraints on the Operating Time of Different Protections: Similarly, the closer to the power source, the longer the operating time of the current protection in the same segment, that is: （9）； In the formula, Set the protection time value for the second stage of the upstream line current. Set the protection time value for the downstream line current stage I or stage II current protection; Considering the protection speed index, take .

5. The method for optimizing the hierarchical protection settings of a distribution network including IIDG and energy storage according to claim 1, characterized in that, Step S2 includes the following steps: S21, Improved Chaotic Initialization: S211, Generating a chaotic Tent sequence: The mathematical expression for the Tent mapping is shown in equation (10). Considering the chaotic characteristics, its control parameters... Get 2.0, input value The output sequence has uniform traversal. Properties of intervals: （10）； In the formula, Let be the chaotic value in the m-th iteration; The chaotic value in the (m+1)th iteration; for two types of optimization variables , Two independent N-dimensional Tent sequences are generated as shown in equation (11), where N is the population size. This represents the m-th iteration chaotic value of the current setting of the j-th segment of the b-th protection. This represents the m-th iteration chaotic value of the b-th protection setting value during the j-th time period; （11）； S212, Mapping the chaotic sequence to the feasible region: As shown in equation (12), the Tent sequence is mapped to the actual search space of the optimization variables; （12）； In the formula, , The current setpoint and action time limit are given for the m-th particle; After mapping, it is necessary to further verify whether the particles satisfy the constraints. Infeasible solutions are corrected by chaotic perturbation to ensure the feasibility of the initial population. S22, Adaptive Splitting Strategy Design: S221, Solution density evaluation: Dynamically replenish sparse region particles based on the density of archived solutions; splitting criterion is: （13）； In the formula, , For the density variance in the archived solution, Density index; For non-dominated solutions in the external archive, calculate the k-nearest neighbor crowding distance for each solution. To balance computational load and accuracy, The density index is shown in equation (14): （14）； In the formula, to avoid the denominator being 0, take... ; S222, Sparse Region Splitting: For a region that satisfies the splitting condition, select the n solutions with the lowest density as splitting centers, where n is one-fifth of the archive size. For each center solution... This generates two new particles, namely: （15）； In the formula, , .

6. The method for optimizing the hierarchical protection settings of a distribution network including IIDG and energy storage according to claim 1, characterized in that: In step S3, the k-means clustering algorithm is introduced to perform cluster analysis on the Pareto solution set, and the optimal solution scheme is determined based on the cluster centers. The specific algorithm flow for finding the optimal solution is as follows: After starting the algorithm process, set various parameters for algorithm operation, including particle swarm size, maximum number of iterations, particle velocity range, learning factor, inertia weight, and density threshold. Next, the Tent chaotic map is used to generate the initial particle swarm. The Tent map generates a chaotic sequence through iteration, and then maps the chaotic sequence to the feasible solution domain of the problem to obtain the initial particle positions, ensuring that the initial population is uniform and comprehensive within the feasible domain. The next step is to check the constraints of each particle in the initial population to ensure that its position meets the constraints of the problem, and to correct the position of particles that do not meet the constraints. The next step is to evaluate the fitness of the corrected particles by calculating the fitness value of each particle under the objective function in order to assess the quality of the solution. Next, an external archive is created and initialized to store the currently found non-dominated solutions, i.e., Pareto optimal solutions; after initialization, the main iteration loop begins. In the main iteration loop, the iteration count is checked first to determine if the current iteration count has reached the maximum value. If it has reached the maximum value, the loop is exited and the Pareto optimal solution set output step is executed; otherwise, the iteration loop step is continued. In the next iteration, the historical best solution for each particle is updated based on the current fitness value, while a global best solution is selected from the external archive. In the next step of the iterative loop, the updated particle position is constrained and corrected if necessary to ensure that the particle is always within the feasible region. The next step in the iterative loop is to evaluate the fitness of the corrected particles, calculate the fitness value of each particle under the objective function, and evaluate the quality of the solution. In the next step of the iterative loop, the new particle is compared with the solution in the external archive, non-dominated sorting and crowding calculation are performed, and the archive is updated to retain the current non-dominated solution; if the archive size exceeds the limit, it is pruned according to the solution density, and solutions in dense regions are removed. The next step in the iterative loop is to calculate the density of solutions in the external archive. The density value reflects the sparsity of the solution distribution in the target space. In the next step of the iterative loop, a density condition judgment is performed to check whether the density variance of the archived solution is greater than a preset threshold. If the condition is met, the following sub-step is executed; if the condition is not met, the third sub-step is executed directly. The first sub-step is to select the solution with the lowest density from the archive, that is, the solution in the sparsest region. The second sub-step involves generating new particles centered on the selected solution using a specific strategy and adding them to the population to enhance the exploration of sparse regions. The third sub-step involves comparing the newly generated particles with the solutions in the archive to update the external archive. In the next iteration, after completing the sub-step, update the individual particle optimality and the global optimality, and return to determine whether the number of iterations has reached the maximum value; Next, after the iteration loop ends, the Pareto optimal solution set stored in the external archive is output as the final result of the algorithm. Next, the algorithm terminates.

7. The method for optimizing the hierarchical protection settings of a distribution network including IIDG and energy storage according to claim 1, characterized in that, It also includes step S4, which involves building an adaptive setpoint strategy library, with the following steps: S41, Define typical operating scenarios: Based on the typical output modes of IIDG and energy storage, define a set of representative typical operating scenarios. This scenario set should cover the main characteristics of system operation, including Peak power generation Charging overnight; S42, Offline generation of strategy library: For each typical scenario, repeat steps S1, S2 and S3 to obtain the optimal protection setting scheme corresponding to the scenario. The schemes of all typical scenarios constitute the "adaptive protection setting strategy library". S43, Online Real-time Adjustment: The system obtains the operating status of the distribution network in real time through the monitoring system, matches the current status with the scenarios in the strategy library, calls the corresponding optimal setpoint, and sends the called new setpoint to each protection device through the communication system, so as to realize the online adaptive adjustment of the protection setpoint.

8. The method for optimizing the hierarchical protection settings of a distribution network including IIDG and energy storage according to claim 7, characterized in that, Step S43 includes the following steps: S431: During normal operation, the distribution network operation status is obtained, and the remote communication device based on MMS transmits the distribution network operation information to the dispatching terminal. S432, calculate IIDG and / or energy storage. The protection setting corresponding to the range, if the fluctuation range of IIDG and / or energy storage output obtained by the dispatching terminal is within... If it is within the range, then no adjustment of the set value is required; If the IIDG and / or energy storage output fluctuations obtained by the dispatcher exceed the allowable output fluctuation range, proceed to the next step to update the set values. S433, Setting update process: The scheduling terminal matches the current running status with typical scenarios in the "Adaptive Protection Setting Strategy Library", calls the corresponding optimal protection setting scheme, and changes the output fluctuation requirements; S434: The new setting is transmitted to the protection device through the communication device and compared with the existing setting of the protection device. If the difference exceeds the allowable range, the protection device stores and uses the adjusted setting, realizing the online update of the protection setting.

Images