Network construction energy storage control parameter identification method based on improved eagle optimization algorithm

Abstract

The invention provides a network construction energy storage control parameter identification method based on an improved eagle optimization algorithm, and belongs to the technical field of power system optimization control and identification. Comprising the following steps of: constructing a network energy storage model comprising a network energy storage battery module, a converter module and a virtual synchronous generator VSG control module according to the specific information collection condition of an actual energy storage system; collecting operation data of the network construction type energy storage under various working conditions to form a data set; to-be-identified control parameters in the network construction energy storage model are selected, relevance of all kinds of control parameters is calculated through a distance correlation coefficient method, and the identification sequence of all kinds of to-be-identified parameters is determined; and constructing an objective function about an actual model and an identification model by using a root-mean-square error, and carrying out step-by-step identification on the control parameters by adopting an improved eagle optimization algorithm according to an identification sequence of the parameters to be identified.

Description

Technical Field

[0001] The present invention relates to the technical field of power system optimization control and identification, and in particular to a method for identifying grid energy storage control parameters based on an improved Osprey optimization algorithm. Background Art

[0002] Grid-connected energy storage systems are a new type of energy storage technology capable of actively supporting grid voltage and frequency, adapting to extremely weak grid operation scenarios. They are a key technical support for building a new power system based primarily on renewable energy. The grid connection process causes the proportion of synchronous generators in the power system to decrease, while the proportion of power electronic equipment increases. Power electronic equipment, with its low inertia and weak damping characteristics, cannot provide the inertia and damping support that synchronous generators provide to the grid. When the system is disturbed, such as by a sudden load change or a fault, the grid frequency can fluctuate dramatically, even leading to system instability.

[0003] As a new type of energy storage technology, grid-connected energy storage systems have the ability to actively support grid voltage and frequency, adapt to extremely weak grid operation scenarios, and are one of the important technical guarantees for building a new power system based on renewable energy. Accurate identification of grid-connected energy storage control parameters is crucial to ensuring the stable operation of the energy storage system in the event of a grid fault. By accurately identifying the control parameters of grid-connected energy storage, the control strategy of the energy storage system can be optimized, and the response speed and regulation accuracy of the energy storage system to grid disturbances can be improved. This allows the system to quickly and accurately compensate for power imbalances in the grid, effectively suppress fluctuations in grid frequency and voltage, and ensure system stability. However, due to the large number of energy storage equipment manufacturers in my country, energy storage equipment from different manufacturers and models has different structures and significant differences in grid-connected characteristics. Furthermore, due to manufacturer confidentiality reasons, the accuracy of the control parameters cannot be guaranteed. Therefore, it is very necessary to propose a grid-connected energy storage control parameter identification method with high identification accuracy and strong adaptability to multiple operating conditions. Summary of the Invention

[0004] In view of this, the present invention provides a grid-type energy storage control parameter identification method based on the improved Osprey optimization algorithm, which is used to improve the accuracy of control parameters of grid-type energy storage modeling and solve the problem of inaccurate grid-type energy storage simulation data.

[0005] The technical solution adopted by the embodiment of the present invention to solve the technical problem is:

[0006] A method for identifying control parameters of grid-connected energy storage based on an improved Osprey optimization algorithm includes:

[0007] Step S1: Building a grid-type energy storage model including a grid-type energy storage battery module, a converter module, and a virtual synchronous generator (VSG) control module according to the specific capital situation of the actual energy storage system;

[0008] Step S2, collecting operating data of the grid-type energy storage under various working conditions to form a data set;

[0009] Step S3: selecting control parameters to be identified in the grid energy storage model, calculating the correlation between various control parameters using the distance correlation coefficient method, and determining the identification order of each type of parameters to be identified;

[0010] In step S4, the objective function of the actual model and the identification model is constructed using the root mean square error, and the control parameters are identified step by step using the improved Osprey optimization algorithm according to the identification order of the parameters to be identified.

[0011] Preferably, the step S1 includes:

[0012] Step S11, building an equivalent circuit model that simulates the dynamic behavior of the energy storage battery module, simulating the changes in the battery's electromotive force, internal resistance, and state of charge;

[0013] Step S12, establishing a mathematical model of a three-phase bridge voltage source energy storage converter, introducing a virtual synchronous machine control to simulate the speed regulation characteristics and excitation characteristics of the synchronous generator, and establishing a control structure of the converter consisting of a power outer loop and a voltage and current inner loop;

[0014] Step S13, using the rotor motion equation to simulate the speed regulation characteristics of the synchronous generator, using the damping coefficient to simulate the synchronous generator damping winding, and establishing the VSG rotor motion equation based on the synchronous generator second-order model;

[0015]

[0016] Where, J v is the virtual inertia of VSG; D v is the virtual damping of VSG; δ is the power angle; P m is the mechanical power; P e is the actual output power of VSG; ω is the VSG output angular frequency; ω0 is the VSG reference angular frequency, ω0 = 2πf, f = 50Hz; Δω is the angular frequency difference;

[0017] Step S14, simulating the excitation characteristics of the synchronous generator through the reactive power-voltage droop control of the VSG, and establishing the reactive power-voltage droop control equation of the VSG:

[0018] U m =U ref +k q (Q ref -Q)

[0019] Where: U m Indicates the calculated value of the terminal voltage; U ref Indicates the terminal voltage reference value; k qIndicates the reactive power droop coefficient; Q ref It represents the reactive power reference value, and Q represents the system output reactive power.

[0020] Preferably, the step S2 includes:

[0021] Step S21: collect the operating data of the grid-type energy storage under various working conditions, calculate the mean μ, standard deviation σ of the data, and the data Z after normalization. i :

[0022]

[0023]

[0024] Where, X i is the original data, n is the total number of samples;

[0025] Step S22: replace outliers and fill missing values ​​with the mean μ, where values ​​with an absolute difference greater than 3σ from the mean μ are considered outliers;

[0026] Step S24: normalize the preprocessed data set.

[0027]

[0028] Where, X n is the normalized data; X min and X max are the minimum and maximum values ​​of the data.

[0029] Preferably, the step S3 includes:

[0030] Step S31, dividing the control parameters to be identified in the grid energy storage model into an outer loop control parameter set and an inner loop control parameter set, wherein the outer loop control parameter set includes virtual inertia, virtual damping, and reactive power droop coefficient, and the inner loop control parameter set includes PI parameters of voltage and current;

[0031] Step S32: For the elements in the same control parameter set, calculate the correlation coefficient dcor(x,y) between any two elements x and y:

[0032]

[0033] Where dcov represents the distance covariance, and the range of the distance correlation coefficient is [0,1], where 0 means that the two variables are completely independent and 1 means that the two variables are completely correlated.

[0034] The sum of the correlation coefficients of each parameter to be identified is calculated, and the sum values ​​are arranged in ascending order as the identification order of the parameters to be identified in the set.

[0035] Preferably, the step S4 includes:

[0036] First, the inner loop control parameters are sequentially identified:

[0037] Step S41: According to the identification order of the parameters to be identified in the inner loop control parameter set, the unknown quantities of the first round of parameter identification are selected, and the objective function J is constructed using the root mean square error:

[0038]

[0039] Where P(k) and Q(k) are the active power and reactive power output in the actual model of the grid-type energy storage, respectively; P*(k) and Q*(k) are the active power and reactive power output in the identification model, respectively; N is the total length of the data;

[0040] Step S42, set the population size to N, the maximum number of iterations to T, and use the chaotic map to generate the initial population; initialize the osprey population x N,m Denoted as:

[0041] x N,m =l max +(l min -l max )×c N,m

[0042] Where: c N,m is the chaos number, N is the number of osprey populations, and m is the dimension of solution; N,m is the initial position vector of the osprey population, l max 、l min are the upper and lower limits of the control parameters to be identified respectively; the matrix form of the osprey population is recorded as:

[0043]

[0044] Where X is the overall matrix of the osprey position, X i is the initial position of the i-th osprey, x i,j is its j-th dimension;

[0045] Step S43, calculating the fitness value of each individual osprey in the osprey population, where each osprey represents a control parameter; the fitness value formula F of the objective function is recorded as:

[0046]

[0047] Where F is the fitness value of the objective function, F(X i) is the fitness value of the i-th osprey, and N is the number of osprey populations;

[0048] Step S44: In the global search phase and the local search phase, the Osprey searches for the global optimal solution through long-distance Weibull random perturbations and searches for the local optimal solution through short-distance Weibull random perturbations. The long-distance step size s1 and the short-distance step size s2 in the Weibull perturbations correspond to the Osprey's position update formula, which is expressed as:

[0049]

[0050] Where: Indicates the updated position of the osprey under the long distance step s1 and short distance step s2 search, SF i represents a position randomly selected from the set of osprey positions with lower objective function values, I i represents the random selection coefficient, wblrnd(α,β,[1,m]) represents the Weibull distribution random function matrix, α and β are the scale and shape parameters of the Weibull perturbation function respectively;

[0051] In the early stage of iteration, a long step size s1 is selected to facilitate global search in the initial stage; in the late stage of iteration, a short step size s2 is selected to facilitate local search in the late stage. The expression is as follows:

[0052]

[0053] Where: η0 is the initial global search step, t is the current iteration number, T is the total number of iterations; η1 is the initial local search step, sign(x) is the sign function, when x is positive, sign(x) is 1, when x is negative, sign(x) is -1; rand(1,h) is a randomly generated numerical matrix;

[0054] Step S45, compare the fitness values ​​of the objective function before and after the osprey position is updated. If the fitness value corresponding to the position before the osprey is updated is better, the osprey maintains the position before the update; if the objective function value corresponding to the updated position is better, the osprey's position is updated. The expression is:

[0055]

[0056] Step S46: Determine the iteration termination condition. If the current number of iterations t reaches the maximum number of iterations T, stop the iteration and output the optimal solution. Otherwise, increase the number of iterations t = t + 1, return to step 4.3, and continue the next iteration until the first round of parameter identification is completed.

[0057] Step S47: Fix the result of the first round of parameter identification as the known quantity, select the unknown quantity for the next round of parameter identification in order, perform identification according to the process of steps S41-S46 and obtain the identification result, and perform parameter identification in sequence until all parameters in the set are identified;

[0058] After the inner loop control parameter sequence identification is completed, the outer loop control parameter sequence identification is performed according to the process of steps S41-S47.

[0059] It can be seen from the above technical solution that the method for identifying the control parameters of grid-type energy storage based on the improved Osprey optimization algorithm provided by the embodiment of the present invention first builds a grid-type energy storage model including a grid-type energy storage battery module, a converter module and a virtual synchronous generator VSG control module according to the specific capital collection situation of the actual energy storage system; collects the operating data of the grid-type energy storage under various working conditions to form a data set; selects the control parameters to be identified in the grid-type energy storage model, calculates the correlation of various control parameters using the distance correlation coefficient method, and determines the identification order of each type of parameter to be identified; constructs the objective function of the actual model and the identification model using the root mean square error, and uses the improved Osprey optimization algorithm to identify the control parameters step by step according to the identification order of the parameters to be identified. The present invention identifies the control parameters in the grid-type energy storage modeling step by step based on the actual operating data of various working conditions, improves the parameter identification efficiency and model accuracy of the grid-type energy storage modeling, and thus provides a reliable reference basis for troubleshooting grid-type energy storage. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 This is a flow chart of identifying control parameters for network energy storage in the present invention;

[0061] Figure 2 It is a block diagram of the parameter identification steps of the improved Osprey optimization algorithm in the present invention;

[0062] Figure 3 1 is a comparison chart of simulation results and test results of working condition 1 in an embodiment of the present invention;

[0063] Figure 4 2 is a comparison chart of simulation results and test results of working condition 2 in an embodiment of the present invention;

[0064] Figure 5 1 is a comparison chart of simulation results and test results for working condition 3 in an embodiment of the present invention;

[0065] Figure 6 1 is a mean square error iteration curve diagram in an embodiment of the present invention. DETAILED DESCRIPTION

[0066] The technical solutions and technical effects of the present invention are further described in detail below with reference to the accompanying drawings of the present invention.

[0067] like Figure 1 As shown, a method for identifying control parameters of low voltage ride-through of grid-connected energy storage based on the improved Osprey optimization algorithm includes the following steps:

[0068] A method for identifying control parameters of grid-connected energy storage based on an improved Osprey optimization algorithm includes:

[0069] Step S1: Based on the specific capital situation of the actual energy storage system, a grid-type energy storage model including a grid-type energy storage battery module, a converter module, and a virtual synchronous generator (VSG) control module is constructed; parameters must be strictly confirmed during the data collection process;

[0070] Step S2, collecting the operating data of the grid-type energy storage under various working conditions to form a data set, and processing the abnormal values ​​and missing values ​​in the original data;

[0071] Step S3: selecting control parameters to be identified in the grid energy storage model, calculating the correlation between various control parameters using the distance correlation coefficient method, and determining the identification order of each type of parameters to be identified;

[0072] In step S4, the objective function of the actual model and the identification model is constructed using the root mean square error, and the control parameters are identified step by step using the improved Osprey optimization algorithm according to the identification order of the parameters to be identified.

[0073] Furthermore, step S1 includes:

[0074] Step S11, building an equivalent circuit model that simulates the dynamic behavior of the energy storage battery module, selecting components such as an ideal voltage source, a series resistor, and a capacitor to form the energy storage battery model, and simulating changes in the battery's electromotive force, internal resistance, and state of charge;

[0075] Step S12, establishing a mathematical model of a three-phase bridge voltage source energy storage converter, introducing a virtual synchronous machine control to simulate the speed regulation characteristics and excitation characteristics of the synchronous generator, and establishing a control structure of the converter consisting of a power outer loop and a voltage and current inner loop;

[0076] Step S13, using the rotor motion equation to simulate the speed regulation characteristics of the synchronous generator, using the damping coefficient to simulate the synchronous generator damping winding, and establishing the VSG rotor motion equation based on the synchronous generator second-order model;

[0077]

[0078] Where, J v is the virtual inertia of VSG; D v is the virtual damping of VSG; δ is the power angle; P m is the mechanical power; P eis the actual output power of VSG; ω is the VSG output angular frequency; ω0 is the VSG reference angular frequency, ω0 = 2πf, f = 50Hz; Δω is the angular frequency difference;

[0079] Step S14, simulating the excitation characteristics of the synchronous generator through the reactive power-voltage droop control of the VSG, and establishing the reactive power-voltage droop control equation of the VSG:

[0080] U m =U ref +k q (Q ref -Q) (2)

[0081] Where: U m Indicates the calculated value of the terminal voltage; U ref Indicates the terminal voltage reference value; k q Indicates the reactive power droop coefficient; Q ref It represents the reactive power reference value, and Q represents the system output reactive power.

[0082] Furthermore, the specific implementation of step S2 includes:

[0083] Step S21: collect the operating data of the grid-type energy storage under various working conditions, calculate the mean μ, standard deviation σ of the data, and the data Z after normalization. i :

[0084]

[0085]

[0086] Where, X i is the original data, and n is the total number of samples. The operating condition generally refers to the voltage fault ride-through condition, including the low-voltage fault ride-through condition and the high-voltage fault ride-through condition. Different operating conditions are set for each type based on five specific factors, as shown in Table 1. The types of operating data include but are not limited to DC voltage, d-axis current, q-axis current, reactive power, and active power:

[0087] Step S22: replace outliers and fill missing values ​​with the mean μ, where values ​​with an absolute difference greater than 3σ from the mean μ are considered outliers;

[0088] Step S24: normalize the preprocessed data set to ensure that each data feature is on the same scale.

[0089]

[0090] Where, X n is the normalized data; X min and X maxare the minimum and maximum values ​​of the data.

[0091] Furthermore, the specific implementation of step S3 includes:

[0092] Step S31: Divide the control parameters to be identified in the grid energy storage model into an outer loop control parameter set and an inner loop control parameter set. The outer loop control parameter set includes virtual inertia, virtual damping, and reactive power droop coefficient, and the inner loop control parameter set includes PI parameters of voltage and current.

[0093] Step S32: For the elements in the same control parameter set, calculate the correlation coefficient dcor(x,y) between any two elements x and y:

[0094]

[0095] Where dcov represents the distance covariance, and the range of the distance correlation coefficient is [0,1], where 0 means that the two variables are completely independent, and 1 means that the two variables are completely correlated. If the distance correlation coefficient is close to 0, it can be considered that there is no significant correlation between the two variables.

[0096] The sum of the correlation coefficients of each parameter to be identified is calculated, and the sum values ​​are arranged in ascending order as the identification order of the parameters to be identified in the set.

[0097] Preferably, step S4 includes:

[0098] First, the inner loop control parameters are sequentially identified:

[0099] Step S41: According to the identification order of the parameters to be identified in the inner loop control parameter set, the unknown quantities of the first round of parameter identification are selected, and the objective function J is constructed using the root mean square error:

[0100]

[0101] Where P(k) and Q(k) are the active power and reactive power output in the actual model of the grid-type energy storage, respectively; P*(k) and Q*(k) are the active power and reactive power output in the identification model, respectively; N is the total length of the data (here refers to the total number of sampling points).

[0102] Step S42, set the population size to N, the maximum number of iterations to T, and use the chaotic map to generate the initial population; initialize the osprey population x N,m Denoted as:

[0103] x N,m =l max +(l min -l max )×c N,m (8)

[0104] Where: c N,m is the chaos number, N is the number of osprey populations, and m is the dimension of solution; N,m is the initial position vector of the osprey population, l max 、l min are the upper and lower limits of the control parameters to be identified respectively; the matrix form of the osprey population is recorded as:

[0105]

[0106] Where X is the overall matrix of the osprey position, X i is the initial position of the i-th osprey, x i,j is its j-th dimension;

[0107] Step S43, calculating the fitness value of each individual osprey in the osprey population, where each osprey represents a control parameter; the fitness value formula F of the objective function is recorded as:

[0108]

[0109] Where F is the fitness value of the objective function, F(X i ) is the fitness value of the i-th osprey, and N is the number of osprey populations;

[0110] Step S44: In the global search phase and the local search phase, the Osprey searches for the global optimal solution through long-distance Weibull random perturbations and searches for the local optimal solution through short-distance Weibull random perturbations. The long-distance step size s1 and the short-distance step size s2 in the Weibull perturbations correspond to the Osprey's position update formula, which is expressed as:

[0111]

[0112] Where: Indicates the updated position of the osprey under the long distance step s1 and short distance step s2 search, SF i represents a position randomly selected from the set of osprey positions with lower objective function values, I i represents the random selection coefficient, wblrnd(α,β,[1,m]) represents the Weibull distribution random function matrix, α and β are the scale and shape parameters of the Weibull perturbation function respectively; m represents the dimension, and t is the current number of iterations;

[0113] In the early stage of iteration, a long step size s1 is selected to facilitate global search in the initial stage; in the late stage of iteration, a short step size s2 is selected to facilitate local search in the late stage. The expression is as follows:

[0114]

[0115] Where: η0 is the initial global search step, t is the current iteration number, T is the total number of iterations; η1 is the initial local search step, sign(x) is the sign function, when x is positive, sign(x) is 1, when x is negative, sign(x) is -1; rand(1,h) is a randomly generated numerical matrix;

[0116] Step S45, compare the fitness values ​​of the objective function before and after the osprey position is updated. If the fitness value corresponding to the position before the osprey is updated is better, the osprey maintains the position before the update; if the objective function value corresponding to the updated position is better, the osprey's position is updated. The expression is:

[0117]

[0118] Step S46: Determine the iteration termination condition. If the current number of iterations t reaches the maximum number of iterations T, stop the iteration and output the optimal solution. Otherwise, increase the number of iterations t = t + 1, return to step 4.3, and continue the next iteration until the first round of parameter identification is completed.

[0119] Step S47: Fix the result of the first round of parameter identification as the known quantity, select the unknown quantity for the next round of parameter identification in order, perform identification according to the process of steps S41-S46 and obtain the identification result, and perform parameter identification in sequence until all parameters in the set are identified;

[0120] After the inner loop control parameter sequence identification is completed, the outer loop control parameter sequence identification is performed according to the process of steps S41-S47.

[0121] The inner-loop control parameters are placed before the outer-loop control parameters because the inner-loop control parameter identification process takes less time. Existing algorithms for parameter tuning typically calculate a group of parameters individually, which can easily lead to local optima and high complexity. Here, the inner-loop control parameters are first classified according to the inner-loop and outer-loop control parameters. Each category of identified parameters is then sorted and identified one by one. Identification can be performed in order from low correlation (i.e., lower correlation) to high correlation, improving identification efficiency, reducing computational effort, and avoiding local optima.

[0122] An embodiment is given below:

[0123] In order to verify the effectiveness of the proposed improved parameter identification method, three groups of low voltage ride-through operating conditions of the grid energy storage are set as shown in Table 1. Three groups of three-phase symmetrical fault conditions are selected, and the initial power is set to 0.9P n , the voltage drop is set to 0.20U n , 0.50U n , 0.90U n, the power factor is set to 0.9, the active power drops are set to 0.20pu, 0.60pu, 0.85pu, and the reactive power drops are set to 2.00pu, 1.00pu, 0.15pu

[0124] Table 1 Three low voltage ride-through conditions

[0125]

[0126] Based on the test data of the actual grid energy storage model under various low voltage ride-through conditions, the proposed identification method is used to identify the control parameters of the grid energy storage. The identification parameters are substituted into the established identification model to complete the simulation test under the same working conditions and obtain the response curve. Finally, the response curve obtained by simulation is compared with the test results of the test data for verification. Figures 3 to 5 The figure is a comparison between the response curve obtained by simulation and the test results obtained by test data. It can be seen that the response curve has a high degree of fit with the test data curve, which illustrates the effectiveness of the proposed method.

[0127] In order to prove the superiority of the improved Osprey optimization algorithm, the Grey Wolf Optimizer (GWO), Genetic Algorithm (GA) and the Improved Osprey Optimization Algorithm (IOA) were compared. The same working conditions were used for testing, and the number of iterations was selected as 25. The objective function convergence curves of the three different optimization algorithms when solving the control parameters were obtained, as shown in the figure. Figure 6 As shown in Figure 2, the initial value of the improved Osprey optimization algorithm is the smallest. During the search and optimization process, the target value found is the smallest. It has significant advantages over the other two algorithms in terms of optimization accuracy and convergence speed.

[0128] Compared with the prior art, the present invention has the following beneficial effects:

[0129] 1. The raw data of the present invention is obtained based on the actual grid energy storage model running under a variety of different working conditions, and the abnormal values ​​and missing values ​​in the raw data are processed, providing a good data basis for the accuracy of subsequent control parameter identification.

[0130] 2. The present invention uses the distance correlation coefficient method to calculate the correlation of various control parameters, determines the identification order of various control parameters, and prioritizes the identification of key parameters with strong correlation and greater impact on the system, thereby improving the efficiency of subsequent control parameter identification.

[0131] 3. The present invention uses chaotic mapping to improve the global search capability of the algorithm in the initial stage. During the algorithm iteration process, the Weibull random perturbation strategy is introduced to adjust the search step size in different iteration periods. This strategy not only helps to prevent the algorithm from falling into local optimality, but also reduces the optimization range in the later stage of the algorithm, thereby improving the accuracy and efficiency of parameter identification.

[0132] The above embodiments are merely preferred technical solutions of the present invention and should not be construed as limiting the present invention. The scope of protection of the present invention shall be the technical solutions set forth in the claims, including equivalent alternatives to the technical features of the technical solutions set forth in the claims. In other words, equivalent alternatives and improvements within this scope are also within the scope of protection of the present invention.

Claims

1. A method for identifying control parameters of grid-connected energy storage based on an improved Osprey optimization algorithm, characterized in that: include: Step S1: Building a grid-type energy storage model including a grid-type energy storage battery module, a converter module, and a virtual synchronous generator (VSG) control module according to the specific capital situation of the actual energy storage system; Step S2, collecting actual operating data of the grid-type energy storage under various working conditions to form a data set; Step S3: selecting control parameters to be identified in the grid energy storage model, calculating the correlation between various control parameters using the distance correlation coefficient method, and determining the identification order of each type of parameters to be identified; In step S4, the objective function of the actual model and the identification model is constructed using the root mean square error, and the control parameters are identified step by step using the improved Osprey optimization algorithm according to the identification order of the parameters to be identified.

2. The method for identifying control parameters of network energy storage based on the improved Osprey optimization algorithm according to claim 1, characterized in that: The step S1 comprises: Step S11, building an equivalent circuit model that simulates the dynamic behavior of the energy storage battery module, simulating the changes in the battery's electromotive force, internal resistance, and state of charge; Step S12, establishing a mathematical model of a three-phase bridge voltage source energy storage converter, introducing a virtual synchronous machine control to simulate the speed regulation characteristics and excitation characteristics of the synchronous generator, and establishing a control structure of the converter consisting of a power outer loop and a voltage and current inner loop; Step S13, using the rotor motion equation to simulate the speed regulation characteristics of the synchronous generator, using the damping coefficient to simulate the synchronous generator damping winding, and establishing the VSG rotor motion equation based on the synchronous generator second-order model; Where, J v is the virtual inertia of VSG; D v is the virtual damping of VSG; δ is the power angle; P m is the mechanical power; P e is the actual output power of VSG; ω is the VSG output angular frequency; ω0 is the VSG reference angular frequency, ω0 = 2πf, f = 50Hz; Δω is the angular frequency difference; Step S14, simulating the excitation characteristics of the synchronous generator through the reactive power-voltage droop control of the VSG, and establishing the reactive power-voltage droop control equation of the VSG: U m =U ref +k q (Q ref -Q) Where: U m Indicates the calculated value of the terminal voltage; U ref Indicates the terminal voltage reference value; k q Indicates the reactive power droop coefficient; Q ref It represents the reactive power reference value, and Q represents the system output reactive power.

3. The method for identifying control parameters of network energy storage based on the improved Osprey optimization algorithm according to claim 2, characterized in that: The step S2 comprises: Step S21: collect the operating data of the grid-type energy storage under various working conditions, calculate the mean μ, standard deviation σ of the data, and the data Z after normalization. i : Where, X i is the original data, n is the total number of samples; Step S22: replace outliers and fill missing values ​​with the mean μ, where values ​​with an absolute difference greater than 3σ from the mean μ are considered outliers; Step S24: normalize the preprocessed data set. Where, X n is the normalized data; X min and X max are the minimum and maximum values ​​of the data.

4. The method for identifying control parameters of network energy storage based on the improved Osprey optimization algorithm according to claim 3 is characterized in that: The step S3 comprises: Step S31, dividing the control parameters to be identified in the grid energy storage model into an outer loop control parameter set and an inner loop control parameter set, wherein the outer loop control parameter set includes virtual inertia, virtual damping, and reactive power droop coefficient, and the inner loop control parameter set includes PI parameters of voltage and current; Step S32: For the elements in the same control parameter set, calculate the correlation coefficient dcor(x,y) between any two elements x and y: Where dcov represents the distance covariance, and the range of the distance correlation coefficient is [0,1], where 0 means that the two variables are completely independent and 1 means that the two variables are completely correlated. The sum of the correlation coefficients of each parameter to be identified is calculated, and the sum values ​​are arranged in ascending order as the identification order of the parameters to be identified in the set.

5. The method for identifying control parameters of network energy storage based on the improved Osprey optimization algorithm according to claim 4, characterized in that: The step S4 comprises: First, the inner loop control parameters are sequentially identified: Step S41: According to the identification order of the parameters to be identified in the inner loop control parameter set, the unknown quantities of the first round of parameter identification are selected, and the objective function J is constructed using the root mean square error: Where P(k) and Q(k) are the active power and reactive power output in the actual model of the grid-type energy storage, respectively; P*(k) and Q*(k) are the active power and reactive power output in the identification model, respectively; N is the total length of the data; Step S42, set the population size to N, the maximum number of iterations to T, and use the chaotic map to generate the initial population; initialize the osprey population x N,m Denoted as: x N,m =l max +(l min -L max )×c N,m Where: c N,m is the chaos number, N is the number of osprey populations, and m is the dimension of solution; N,m is the initial position vector of the osprey population, l max 、l min are the upper and lower limits of the control parameters to be identified respectively; the matrix form of the osprey population is recorded as: Where X is the overall matrix of the osprey position, X i is the initial position of the i-th osprey, x i,j is its j-th dimension; Step S43, calculating the fitness value of each individual osprey in the osprey population, where each osprey represents a control parameter; the fitness value formula F of the objective function is recorded as: Where F is the fitness value of the objective function, F(X i ) is the fitness value of the i-th osprey, and N is the number of osprey populations; Step S44: In the global search phase and the local search phase, the Osprey searches for the global optimal solution through long-distance Weibull random perturbations and searches for the local optimal solution through short-distance Weibull random perturbations. The long-distance step size s1 and the short-distance step size s2 in the Weibull perturbations correspond to the Osprey's position update formula, which is expressed as: Where: Indicates the updated position of the osprey under the long distance step s1 and short distance step s2 search, SF i represents a position randomly selected from the set of osprey positions with lower objective function values, I i represents the random selection coefficient, wblrnd(α,β,[1,m]) represents the Weibull distribution random function matrix, α and β are the scale and shape parameters of the Weibull perturbation function respectively; t is the current number of iterations; In the early stage of iteration, a long step size s1 is selected to facilitate global search in the initial stage; in the late stage of iteration, a short step size s2 is selected to facilitate local search in the late stage. The expression is as follows: Where: η0 is the initial global search step, t is the current iteration number, T is the total number of iterations; η1 is the initial local search step, sign(x) is the sign function, when x is positive, sign(x) is 1, when x is negative, sign(x) is -1; rand(1,h) is a randomly generated numerical matrix; Step S45, compare the fitness values ​​of the objective function before and after the osprey position is updated. If the fitness value corresponding to the position before the osprey is updated is better, the osprey maintains the position before the update; if the objective function value corresponding to the updated position is better, the osprey's position is updated. The expression is: Step S46: Determine the iteration termination condition. If the current number of iterations t reaches the maximum number of iterations T, stop the iteration and output the optimal solution. Otherwise, increase the number of iterations t = t + 1, return to step 4.3, and continue the next iteration until the first round of parameter identification is completed. Step S47: Fix the result of the first round of parameter identification as the known quantity, select the unknown quantity for the next round of parameter identification in order, perform identification according to the process of steps S41-S46 and obtain the identification result, and perform parameter identification in sequence until all parameters in the set are identified; After the inner loop control parameter sequence identification is completed, the outer loop control parameter sequence identification is performed according to the process of steps S41-S47.

Images