Energy storage system pi control parameter identification method based on improved grey wolf algorithm

Abstract

A method for identifying PI control parameters of an energy storage system based on an improved gray wolf algorithm is used to establish a measured model of the energy storage system. The grid-side inverter in this measured model employs a dual-closed-loop control method. The measured model is then run to obtain the parameters during steady-state operation. d , q Axis current curves; establish an energy storage system identification model, and determine the parameters to be identified as the PI control parameters of the grid-side inverter based on the measured model of the energy storage system; assign random values ​​to the PI control parameters of the energy storage system identification model, and obtain the output of the energy storage system identification model. d , q The shaft response curves are used to calculate the root mean square error of both; the outputs of the energy storage system identification model and the measured model of the energy storage system are used. d , q The root mean square error of the axial response curve is used as the objective function. An improved Grey Wolf algorithm is employed to find the optimal value of the objective function, and the resulting solution set represents the parameters to be identified. This method provides parameter identification results with higher consistency and faster convergence speed.

Description

Technical Field

[0001] This invention relates to the field of new energy power generation system technology, and in particular to a method for identifying PI control parameters of energy storage systems based on an improved gray wolf algorithm. Background Technology

[0002] In recent years, the grid-connected capacity of new energy sources has grown rapidly, and the proportion of new energy in the power system has also increased significantly. New energy units are connected to the grid via power electronic inverters. These nonlinear power electronic inverters exhibit severe broadband oscillations and large harmonic components, leading to strong uncertainty, low inertia, weak disturbance rejection, and strong nonlinearity in the power grid. Furthermore, because new energy power generation is affected by weather and diurnal variations, its power output exhibits fluctuating and intermittent characteristics, resulting in complex interactions and profound changes in the operating characteristics of the power grid.

[0003] Currently, the main instability factors facing the power grid are caused by the large-scale grid connection of new energy sources. On the one hand, the power output of new energy sources, represented by photovoltaics and wind power, is random, fluctuating, and intermittent. Diverse and severe natural weather further amplifies this uncertainty, contributing to grid instability. In future grids with ultra-high proportions of new energy, the application of technologies such as sufficient new energy power and energy storage, demand-side response, time-series production simulation, and artificial intelligence is expected to solve the power imbalance problem caused by the volatility of new energy sources. On the other hand, new energy converters based on power electronics grid connection technology have replaced a large number of traditional synchronous generator units, fundamentally changing the operating characteristics of the power system and posing significant challenges to the stable operation of the new power grid.

[0004] Establishing an accurate simulation model of the energy storage system is fundamental for analyzing and calculating power systems with large-scale integration of renewable energy units. It is also a prerequisite and key to power system dispatching, analysis of renewable energy unit operating characteristics, and power quality analysis. To study the operation and control issues of power systems containing renewable energy units, correct model equations and accurate model parameters for the renewable energy units are required. However, there are numerous brands and models of actual renewable energy units, and most manufacturers do not provide accurate model parameters due to insufficient testing technology or intellectual property protection reasons. These parameters can only be obtained through measurement identification methods. Based on a hardware-in-the-loop simulation test platform, different operating scenarios are set up for testing, and response curves of different electrical parameters are obtained. Relevant parameters reflecting the steady-state and dynamic characteristics of the actual controller are selected for identification, and a real-world model of the renewable energy system is built to meet the simulation requirements for real-world modeling of renewable energy controllers.

[0005] Existing gray wolf optimization algorithms suffer from low efficiency and accuracy in identifying energy storage system model parameters. The main issues are as follows: For actual controllers, their parameters are unknown, so initial parameter values ​​lack reference values ​​and are typically set based on experience. Inappropriate initial values ​​lead to significant deviations in the final identification results. Traditional basic gray wolf algorithms are prone to getting trapped in local optima when applied to energy storage system model parameter identification, further contributing to inaccurate results. Furthermore, existing gray wolf optimization algorithms exhibit randomness, which can prolong identification time and reduce efficiency during parameter optimization. In conclusion, existing gray wolf optimization algorithms are unsuitable for identifying energy storage system model parameters. Summary of the Invention

[0006] To address the issues of low efficiency and low accuracy in identifying energy storage system model parameters using the existing Grey Wolf optimization algorithm, this invention provides a method for identifying PI control parameters of energy storage systems based on an improved Grey Wolf algorithm. Compared with traditional methods, this method achieves higher consistency in parameter identification results and faster convergence speed.

[0007] The technical solution adopted in this invention is as follows:

[0008] The method for identifying PI control parameters of energy storage systems based on the improved gray wolf algorithm includes the following steps:

[0009] S1: Establish a measured model of the energy storage system. The grid-side inverter of the measured model of the energy storage system adopts a dual closed-loop control method. Run the measured model of the energy storage system and obtain the d-axis and q-axis current curves during steady-state operation as measured response data.

[0010] S2: Establish an energy storage system identification model, and determine the parameters to be identified as the grid-side inverter PI control parameters based on the measured model of the energy storage system.

[0011] S3: Within a certain range, assign random values ​​to the PI control parameters of the energy storage system identification model, obtain the d-axis and q-axis response curves of the energy storage system identification model, and calculate the root mean square error of both.

[0012] S4: The parameter identification problem is transformed into a multi-objective optimization problem. The root mean square error of the d-axis and q-axis response curves output by the energy storage system identification model and the energy storage system measured model is used as the objective function. The improved Grey Wolf Algorithm (CGWO) is used to solve for the optimal value of the objective function. The solution set obtained is the result of the parameters to be identified.

[0013] S5: Substitute the identification results into the energy storage system identification model, run the energy storage system identification model, and obtain the d-axis and q-axis current response curves output by the energy storage system identification model. Compare the d-axis and q-axis current response curves output by the energy storage system identification model with those output by the measured energy storage system model, calculate the error, and evaluate the accuracy of the identification results. S6: Set different voltage drop depths, obtain the d-axis and q-axis current response curves of the energy storage system identification model and the measured energy storage system model under various operating conditions, calculate the error of the response curves under each operating condition, and evaluate the adaptability of the identification results to transient operating conditions.

[0014] S1 includes the following steps:

[0015] S1.1: Set a set of fixed dual closed-loop PI control parameters and fill them into the measured model of the energy storage system;

[0016] S1.2: Run the actual test model of the energy storage system and convert the current in the three-phase rotating coordinate system (a, b, c) into the current in the two-phase stationary coordinate system through d and q transformations.

[0017] S1.3: Collect d-axis and q-axis current data 1 second before the cutoff to enter steady state, and form a feature dataset.

[0018] S2 includes the following steps:

[0019] S2.1: Establish the voltage outer loop controller model. The expression for the voltage outer loop controller model is:

[0020]

[0021] Where: K pd K id These are the proportional and integral coefficients of the DC voltage outer loop controller, respectively; U dc U dc_ref These are the actual and reference values ​​of the DC voltage, respectively; i d_ref i q_ref These are the reference values ​​output by the DC voltage outer loop controller and the grid-connected voltage outer loop controller, respectively.

[0022] S2.2: Establish the current inner loop controller model, the expression of which is:

[0023]

[0024] Among them, K pi K ii These are the proportional and integral coefficients of the inner loop current controller, respectively; u d u q These are the d-axis and q-axis voltage control signals, respectively; U sd Usq S2.3: The proportional coefficient K of the DC voltage outer loop controller in the dual closed-loop controller model of the energy storage system established in S2.1 and S2.2. (The components of grid-connected voltage on the d and q axes are given, respectively; L is the filter inductance; ω is the synchronous angular velocity.) pd and integral coefficient K id The proportional coefficient K of the current inner loop controller pi and integral coefficient K ii The parameter to be identified has been determined.

[0025] In step S3, the d-axis and q-axis response curves output by the measured model of the energy storage system are denoted as i. d_act and i q_act The d-axis and q-axis response curves output by the energy storage system identification model are denoted as i. d_sim and i q_sim The expression for the objective function f is as follows:

[0026]

[0027] S4 includes the following steps:

[0028] S4.1: Treat the objective function f as the object of optimization;

[0029] S4.2: Randomly generate individuals for the initial population in the exploration space, as shown in equation (4):

[0030] x i =rand(1,D).*(ub-lb)+lb (4)

[0031] In the formula: x i Let denot 'D' be the individual position, 'D' be the dimension, and 'ub' and 'lb' be the upper and lower boundaries, respectively.

[0032] S4.3: Calculate the fitness value of all individuals in the population, i.e., the size of f, and select the three individuals with the smallest fitness values, which are denoted as α, β, and γ wolves, respectively.

[0033] S4.4: During the hunting process, when gray wolves surround and hunt prey, the distance between the individual and the prey is as shown in equation (5):

[0034] l = |CX p (t)-X(t)| (5)

[0035] In the formula: l is the distance between the individual and the prey, C is the position control coefficient, t is the current iteration number, and X... p X and X represent the positions of the prey and the gray wolf, respectively. The formula for updating the gray wolf's position is as follows:

[0036] X(t+1)=CX p (t)-Al (6)

[0037] A = 2a·r1-a (7)

[0038] C = 2·r² (8)

[0039] In the formula: A is the position movement control coefficient, a is the convergence factor, and r1 and r2 are both random numbers between [0,1]. The expression for a is shown in equation (9):

[0040]

[0041] In the formula: t is the current iteration number, t max Let a be the maximum number of iterations. initial and a final are the initial and final values ​​of the convergence factor, respectively, and n is the decreasing exponent, ranging from (0,1).

[0042] S4.5: During the hunting process, the mathematical model for a gray wolf tracking its prey is as follows:

[0043]

[0044] In the formula: l α l β l γ X represents the distance of α, β, and γ wolves from other individuals. α X β X γ Let C1, C2, and C3 be the current positions of α, β, and γ wolves, respectively, where C1, C2, and C3 are random values, and X is the current position of the gray wolf. The formula for updating the individual positions in this stage is as follows:

[0045]

[0046]

[0047]

[0048]

[0049] In the formula: W α W β W γ The learning rates are α, β, and γ wolves, respectively.

[0050] S4.6: Based on steps S4.1 to S4.5, solve for the solution that minimizes the objective function f; the solution set is the parameter to be identified. In S5, substitute the identification results into the energy storage system identification model, simulate to obtain the d-axis and q-axis current waveforms, and calculate the error between the simulation results and the measured results, using the following formula:

[0051]

[0052]

[0053] Where: δ d and δ q These are the relative errors of the d-axis and q-axis currents, respectively.

[0054] S6 includes the following steps:

[0055] S6.1: Set the voltage drop to 20% to 80%, in 5% increments;

[0056] S6.2: Substitute the identification results into the energy storage system identification model and simulate to obtain the d-axis and q-axis current waveforms;

[0057] S6.3: Calculate the error between the simulation results and the measured results, using the following formula:

[0058]

[0059]

[0060] This invention provides a method for identifying PI control parameters of an energy storage system based on an improved gray wolf algorithm, with the following technical advantages:

[0061] 1) This invention improves the gray wolf algorithm, evaluates fitness and global optimal solution, iteratively updates the population position until the PI control parameters of the energy storage system that satisfy the global optimal solution are found and output. Compared with traditional methods, the parameter identification results have higher consistency and faster convergence speed.

[0062] 2) This invention identifies PI control parameters for energy storage system models in time-domain simulation, which can effectively reflect the actual operating state of the energy storage system model and its impact on the stable operation of the actual system.

[0063] 3) This invention introduces a dynamic convergence factor and proportional weights to improve the population initialization of the Grey Wolf algorithm, enabling the algorithm to generate initial solutions with good diversity in the search space. High-quality population initialization effectively improves the convergence speed and solution accuracy of the traditional Grey Wolf algorithm. By adjusting the proportional weights to improve the position update formula, the problem of the traditional Grey Wolf algorithm easily getting trapped in local optima is effectively solved.

[0064] 4) This invention uses an improved gray wolf optimization algorithm to identify the PI control parameters of the energy storage system model, which can effectively improve the efficiency and accuracy of parameter identification. Attached Figure Description

[0065] Figure 1 This is a flowchart of a method for identifying PI control parameters of an energy storage system based on an improved gray wolf algorithm, as described in an embodiment of the present invention.

[0066] Figure 2 This is a control block diagram of the dual closed-loop controller model of the energy storage system in an embodiment of the present invention.

[0067] Figure 3 This is a comparison chart of the d-axis current response curve of the energy storage system identification model under steady-state operation and the measured curve in an embodiment of the present invention.

[0068] Figure 4 This is a comparison chart of the q-axis current response curve of the energy storage system identification model under steady-state operation and the measured curve in an embodiment of the present invention.

[0069] Figure 5 This is a schematic diagram of the convergence curve of the identification process of the energy storage system identification model in an embodiment of the present invention. Detailed Implementation

[0070] like Figure 1 As shown, the method for identifying PI control parameters of an energy storage system based on the improved gray wolf algorithm includes the following steps:

[0071] S1: Establish a measured model of the energy storage system. The grid-side inverter of the measured model adopts a dual closed-loop control method. Run the measured model of the energy storage system and obtain the d-axis and q-axis current curves during steady-state operation as measured response data.

[0072] S2: Establish an energy storage system identification model, and determine the parameters to be identified as grid-side inverter PI control parameters based on the measured model of the energy storage system.

[0073] S3: Within a certain range, assign random values ​​to the PI control parameters of the identification model, obtain the d-axis and q-axis response curves of the identification model output, and calculate the root mean square error of both.

[0074] S4: The parameter identification problem is transformed into a multi-objective optimization problem. The root mean square error of the d-axis and q-axis response curves output by the identification model and the measured model is used as the objective function. The improved Grey Wolf Algorithm (CGWO) is used to solve for the optimal value of the objective function. The solution set obtained is the result of the parameter to be identified.

[0075] S5: Substitute the identification results into the energy storage system identification model, run the identification model, obtain the d-axis and q-axis current response curves output by the identification model, compare the d-axis and q-axis current response curves output by the identification model with the d-axis and q-axis current response curves output by the measured model, calculate the error, and evaluate the accuracy of the identification results.

[0076] S6: Set different voltage drop depths, obtain the d-axis and q-axis current response curves of the identified model and the measured model under various operating conditions, calculate the error of the response curves under various operating conditions, and evaluate the adaptability of the identification results to transient operating conditions.

[0077] Specifically, in step S1, a set of fixed dual-closed-loop PI control parameters are first set and filled into the energy storage system test model; the energy storage system test model is run, and the current in the three-phase rotating coordinate system of a, b, and c is converted into the current in the two-phase stationary coordinate system through d and q transformation; the d and q axis current data are collected 1 second before entering steady state to form a feature dataset.

[0078] like Figure 2 The diagram shown is a control block diagram of an energy storage system, where ω is the angular velocity and L is the inductance. Specifically, in step S2, a voltage outer loop controller model is established, and the expression of the voltage outer loop controller model is:

[0079]

[0080] Where: K pd K id These are the proportional and integral coefficients of the DC voltage outer loop controller, U dc U dc_ref These are the actual and reference values ​​of the DC voltage, i d_ref i q_ref The reference values ​​for the DC voltage outer loop controller output and the grid-connected voltage outer loop controller output are respectively provided; a current inner loop controller model is established, and the expression of the current inner loop controller model is as follows:

[0081]

[0082] Among them, K pi K ii These are the proportional and integral coefficients of the inner loop current controller, respectively; u d u q These are the d-axis and q-axis voltage control signals, respectively; U sd U sq These are the d-axis and q-axis components of the grid-connected voltage, respectively; L is the filter inductance; and ω is the synchronous angular velocity.

[0083] The proportional coefficient K of the DC voltage outer loop controller in the dual closed-loop controller model of the energy storage system is... pd and integral coefficient K id The proportional coefficient K of the current inner loop controller pi and integral coefficient K ii The parameter to be identified has been determined.

[0084] Specifically, in step S3, the d-axis and q-axis response curves output by the measured model are denoted as i. d_act and i q_act The d-axis and q-axis response curves output by the identification model are denoted as i. d_sim and i q_simThe expression for the objective function f is as follows:

[0085]

[0086] Specifically, in step S4, the objective function f is used as the object of optimization.

[0087] The individuals of the initial population are randomly generated in the exploration space, as shown in equation (4):

[0088] x i =rand(1,D).*(ub-lb)+lb (4)

[0089] In the formula: x i Let D be the individual position, ub be the dimension, and lb be the upper and lower boundaries, respectively.

[0090] Calculate the fitness value of all individuals in the population, i.e., the size of f, and select the three individuals with the smallest fitness values, which are denoted as α, β, and γ wolves, respectively.

[0091] During the hunting process, when gray wolves surround and hunt their prey, the distance between the individual and the prey is shown in equation (5):

[0092] l = |CX p (t)-X(t)| (5)

[0093] In the formula: l is the distance between the individual and the prey, C is the position control coefficient, t is the current iteration number, and X... p X and X represent the positions of the prey and the gray wolf, respectively. The formula for updating the gray wolf's position is as follows:

[0094] X(t+1)=CX p (t)-AD (6)

[0095] A = 2a·r1-a (7)

[0096] C = 2·r² (8)

[0097] In the formula: A is the position movement control coefficient, a is the convergence factor, and r1 and r2 are both random numbers between [0,1]. The expression for a is shown in equation (9):

[0098]

[0099] In the formula: t is the current iteration number, t max Let a be the maximum number of iterations. initial and a final are the initial and final values ​​of the convergence factor, respectively, and n is the decreasing exponent, ranging from (0,1).

[0100] During a hunt, the mathematical model for a gray wolf tracking its prey is as follows:

[0101]

[0102] In the formula: l α l β l γ X represents the distance of α, β, and γ wolves from other individuals. α X β X γ Let C1, C2, and C3 be the current positions of α, β, and γ wolves, respectively, where C1, C2, and C3 are random values, and X is the current position of the gray wolf. The formula for updating the individual positions in this stage is as follows:

[0103]

[0104]

[0105]

[0106]

[0107] In the formula: W α W β W γ Let be the learning rates for α, β, and γ wolves, respectively. Following the steps above, find the solution that minimizes the objective function f; this solution set represents the parameters to be identified.

[0108] Specifically, in step S5, the identification results are substituted into the energy storage system simulation model to simulate and obtain the d-axis and q-axis current waveforms; the error between the simulation results and the measured results is calculated using the following formula:

[0109]

[0110]

[0111] Where: δ d and δ q These are the relative errors of the d-axis and q-axis currents, respectively.

[0112] Specifically, in step S6, the voltage is set to drop to 20% to 80% in 5% increments; the identification results are substituted into the energy storage system simulation model to obtain the d-axis and q-axis current waveforms; the error between the simulation results and the measured results is calculated using the following formula:

[0113]

[0114]

[0115] Figure 3 A schematic diagram of the convergence curve of the identification process, such as Figure 4 and Figure 5 The figures show the d-axis and q-axis current curves of the energy storage system identification model under steady-state operation. Figure 3 It can be seen that the convergence speed of the identification process is relatively fast, from Figure 4 and Figure 5 It can be seen that the identification results of the d-axis and q-axis currents have small errors compared with the measured results, proving that the identification results have high accuracy.

[0116] This invention identifies parameters of an energy storage system model in time-domain simulation, effectively demonstrating the impact of the model's actual operating state on the stable operation of the actual system. The invention improves the population initialization of the Grey Wolf algorithm by introducing a dynamic convergence factor and proportional weights, enabling the algorithm to generate initial solutions with good diversity in the search space. High-quality population initialization effectively improves the convergence speed and solution accuracy of the traditional Grey Wolf algorithm. Furthermore, by adjusting the proportional weights to improve the position update formula, the problem of the traditional Grey Wolf algorithm easily getting trapped in local optima is effectively solved.

Claims

1. A method for identifying PI control parameters of an energy storage system based on an improved gray wolf algorithm, characterized in that... Includes the following steps: S1: Establish a test model of the energy storage system. The grid-side inverter in the test model adopts a dual-closed-loop control method. Run the test model of the energy storage system to obtain the steady-state operation data. d , q The shaft current curve is used as measured response data. S2: Establish an energy storage system identification model, and determine the parameters to be identified as the grid-side inverter PI control parameters based on the measured model of the energy storage system. S3: Assign random values ​​to the PI control parameters of the energy storage system identification model and obtain the output of the energy storage system identification model. d , q Calculate the root mean square error of both axis response curves; S4: Transform the parameter identification problem into a multi-objective optimization problem, using the outputs of the energy storage system identification model and the measured model of the energy storage system. d , q The root mean square error of the axis response curve is used as the objective function. The improved Grey Wolf Algorithm (CGWO) is used to solve for the optimal value of the objective function, and the solution set obtained is the result of the parameter to be identified. S5: Substitute the identification results into the energy storage system identification model, run the energy storage system identification model, and obtain the output of the energy storage system identification model. d , q The shaft current response curve is the output of the energy storage system identification model. d , q The shaft current response curve and the output of the measured model of the energy storage system d , q The shaft current response curves are compared, the error is calculated, and the accuracy of the identification results is evaluated. S6: Set different voltage drop depths to obtain the energy storage system identification model and the energy storage system measured model under various operating conditions. d , q The shaft current response curve is calculated, the error of the response curve under each operating condition is determined, and the adaptability of the identification results to transient operating conditions is evaluated.

2. The method for identifying PI control parameters of an energy storage system based on the improved gray wolf algorithm according to claim 1, characterized in that: S1 includes the following steps: S1.1: Set a set of fixed dual closed-loop PI control parameters and fill them into the measured model of the energy storage system; S1.2: Run the actual test model of the energy storage system, through d , q Transformation will a , b , c The current in a three-phase rotating coordinate system is converted into the current in a two-phase stationary coordinate system. S1.3: Data acquisition ends 1 second before reaching steady state. d , q The shaft current data constitutes the feature dataset.

3. The method for identifying PI control parameters of an energy storage system based on the improved gray wolf algorithm according to claim 1, characterized in that: S2 includes the following steps: S2.1: Establish the voltage outer loop controller model. The expression for the voltage outer loop controller model is: （1）； In the formula: K pd , K id These are the proportional and integral coefficients of the DC voltage outer loop controller, respectively. U dc , U dc_ref These are the actual and reference values ​​of the DC voltage, respectively. i d_ref , i q_ref These are the reference values ​​output by the DC voltage outer loop controller and the grid-connected voltage outer loop controller, respectively. S2.2: Establish the current inner loop controller model, the expression of which is: （2）； in, K pi , K ii These are the proportional and integral coefficients of the inner loop current controller, respectively. u d , u q These are the d-axis and q-axis voltage control signals, respectively. U sd , U sq These are the d-axis and q-axis components of the grid-connected voltage, respectively. L For filtering inductors, ω Synchronous angular velocity; S2.3: The proportional gain of the DC voltage outer loop controller in the dual closed-loop controller model of the energy storage system established in S2.1 and S2.

2. K pd and integral coefficient K id The proportional gain of the current inner loop controller K pi and integral coefficient K ii The parameter was identified as the control parameter to be identified.

4. The method for identifying PI control parameters of an energy storage system based on the improved gray wolf algorithm according to claim 1, characterized in that: In S3, the measured model of the energy storage system is output. d , q The axis response curves are denoted as follows: i d_act and i q_act The output of the energy storage system identification model d , q The axis response curves are denoted as follows: i d_sim and i q_sim Objective function f The expression is as follows: （3）。 5. The method for identifying PI control parameters of an energy storage system based on the improved gray wolf algorithm according to claim 1, characterized in that: S4 includes the following steps: S4.1: The objective function f As the object of optimization; S4.2: Randomly generate individuals for the initial population in the exploration space, as shown in equation (4): （4）； In the formula: x i For individual position, D As a dimension, ub and lb These are the upper and lower boundaries, respectively; S4.3: Calculate the fitness value of all individuals in the population, i.e. f Based on the size, select the three smallest individuals, and denote the three individuals as follows: α, β, γ Wolf; S4.4: During the hunting process, when gray wolves surround and hunt prey, the distance between the individual and the prey is as shown in equation (5): （5）； In the formula: l The distance between an individual and its prey. C For position control coefficients, t This represents the current iteration number. X p and X These represent the positions of the prey and the gray wolf, respectively. The formula for updating the gray wolf's position is as follows: （6）； （7）； （8）； In the formula: A This is the position movement control coefficient. a The convergence factor is r 1 and r 2 are all random numbers between [0,1], as shown in equation (9). a The expression: （9）； In the formula: t This represents the current iteration number. t max The maximum number of iterations, a initial and a final These are the initial and final values ​​of the convergence factor, respectively. n The exponent is decreasing and its range is (0,1]. S4.5: During the hunting process, the mathematical model for a gray wolf tracking its prey is as follows: （10）； In the formula: l α , l β , l γ They are respectively α, β, γ The wolf's position relative to other individuals, X α , X β , X γ They are respectively α, β, γ The wolf's current location C 1. C 2. C All 3 are random values. X This represents the current position of the gray wolf; the formula for updating the individual's position during this stage is as follows: （11）； （12）； （13）； （14）； In the formula: W α , W β , W γ They are respectively α, β, γ The learning rate of wolves; S4.6: Solve for the objective function based on steps S4.1 to S4.

5. f The smallest solution has a solution set that is the parameter to be identified.

6. The method for identifying PI control parameters of an energy storage system based on the improved gray wolf algorithm according to claim 1, characterized in that: In step S5, the identification results are substituted into the energy storage system identification model, and the d-axis and q-axis current waveforms are obtained through simulation. The error between the simulation results and the measured results is calculated using the following formula: （15）； （16）； In the formula: δ d and δ q These are the relative errors of the d-axis and q-axis currents, respectively. i d_act and i q_act The outputs of the measured model of the energy storage system are respectively d , q Axis response curve; i d_sim and i q_sim The outputs of the energy storage system identification model are respectively d , q Axis response curve.

7. The method for identifying PI control parameters of an energy storage system based on the improved gray wolf algorithm according to claim 1, characterized in that: S6 includes the following steps: S6.1: Set the voltage drop to 20% to 80% in 5% increments; S6.2: Substitute the identification results into the energy storage system identification model and simulate to obtain the d-axis and q-axis current waveforms; S6.3: Calculate the error between the simulation results and the measured results, using the following formula: （17）； （18）； In the formula: i d_act and i q_act The outputs of the measured model of the energy storage system are respectively d , q Axis response curve; i d_sim and i q_sim The outputs of the energy storage system identification model are respectively d , q Axis response curve.

Images