SVG control parameter identification method based on improved grey wolf algorithm

Abstract

The invention discloses an SVG control parameter identification method based on an improved grey wolf algorithm, and the method comprises the following steps: S1, analyzing the control principle of an SVG, building an SVG electromechanical transient model, and determining parameters needing to be identified; s2, hardware-in-loop testing is carried out on an RT-LAB platform, and data sets of an SVG actual controller under different fault working conditions are obtained; s3, on the basis of the HIL data set, identifying SVG control parameters by using an improved grey wolf algorithm, and calculating errors of the SVG control parameters; according to the method, the identification accuracy can be improved, and SVG high-precision modeling can be realized.

Description

Technical Field

[0001] The present invention relates to the field of SVG parameter identification, and particularly to a method for identifying SVG control parameters based on an improved grey wolf algorithm. Background Art

[0002] With the accelerating transformation of the global energy structure towards low-carbon and clean directions, the installed capacity of new energy power generation represented by wind power and photovoltaic power has been continuously climbing. By the end of 2024, the total installed capacity of renewable energy power generation in China had reached 1.889 billion kilowatts, accounting for 56% of the total national installed power generation capacity. This profound change in the energy structure not only injects green kinetic energy into the power system but also brings significant technical challenges. A large number of distributed new energy sources are connected to the power grid through power electronic conversion devices, which fundamentally changes the operating characteristics of the system, resulting in increasingly prominent power quality problems such as voltage fluctuations, harmonic pollution, and three-phase imbalance, seriously threatening the safe and stable operation of the power system. Against this background, as the core equipment of the flexible AC transmission system, SVG plays a key role in improving the power quality of the power grid by virtue of its technical advantages of fast dynamic response speed and high compensation accuracy. However, the interaction effects generated after SVG is connected to the power grid exhibit significant time-varying and non-linear characteristics. Existing research shows that the mismatch of SVG controller parameters will lead to the deterioration of the system damping characteristics and may induce new stability problems such as subsynchronous oscillation under specific working conditions. Therefore, studying the SVG parameter identification method has important theoretical value and engineering significance. By establishing an accurate SVG model and developing a parameter identification algorithm based on measured data, not only can the accuracy of power system simulation analysis be improved, but also a theoretical basis can be provided for the optimization of SVG controller parameters. This will effectively enhance the grid connection adaptability of new energy power stations and is of strategic significance for building a new power system and ensuring the safety of the power grid under the access of a high proportion of renewable energy. However, there is currently little research on SVG parameter identification, and the identification accuracy is generally not high. Therefore, how to achieve high-precision SVG parameter identification is a key problem that urgently needs to be solved in power system simulation analysis. Summary of the Invention

[0003] The purpose of the present invention is to overcome the above deficiencies and provide a method for identifying SVG control parameters based on an improved grey wolf algorithm to improve the accuracy of identification and achieve high-precision modeling of SVG.

[0004] To solve the above technical problems, the technical solution adopted by the present invention is: A method for identifying SVG control parameters based on an improved grey wolf algorithm, comprising the following steps:

[0005] Step S1, analyze the control principle of SVG and establish an electromechanical transient model of SVG to clarify the parameters to be identified;

[0006] Step S2, conduct hardware-in-the-loop testing on the RT-LAB platform to obtain the dataset of the actual SVG controller under different fault conditions;

[0007] Step S3, based on the HIL dataset, use the improved grey wolf algorithm to identify the SVG control parameters and calculate their errors.

[0008] Preferably, the said Step S1 includes:

[0009] Step S11, establish an SVG electromechanical model according to the SVG control strategy in the PSASP Power System Analysis Comprehensive Program;

[0010] Step S12, based on the control strategy in S11, the differential equation of the SVG controller is as follows:

[0011]

[0012] In the formula, V T1 is the output of the voltage measurement link, V T23 and V T45 are the outputs of the first-stage and second-stage lead-lag links respectively, V p and V i are the outputs of the proportional link and the integral link respectively; U is the voltage at the SVG output; U ref is the voltage reference value; U S is the AC side voltage; I S is the SVG output current; T1 is the time constant of the measurement loop; T2, T3, T4, T5 are the four different lead-lag time constants in the PI regulator; T P is the proportional link time constant; T S is the trigger delay; K p , K i are the proportional and integral coefficients of the PI controller respectively; K D is the slope of the response characteristic curve; L is the filter reactance between the SVG and the system, which is a known quantity; U max , U min are the upper and lower limits of the voltage; I Cmax , I Lmax are the maximum capacitive and inductive currents; U smax , U smin are the limits of the proportional and integral links;

[0013] Step S13, based on the research in Step S12, the parameter set to be identified for the SVG controller model is:

[0014]

[0015] Preferably, the said Step S2 includes:

[0016] Step S21, build a simulation system required for HIL testing;

[0017] Step S22, HIL data acquisition: Based on the simulation system built in Step S21, connect RT-LAB to the actual SVG controller through optical fiber, test the fault response characteristics of the SVG under various fault conditions, and generate corresponding simulation data sets.

[0018] Step S23, define the fitness function.

[0019] Preferably, in Step S21, the simulation system consists of a 220 kV infinite power source, a transformer with a rated capacity of 100 MVA and a rated transformation voltage level of 220 / 35 kV, a filter inductor L, and a to-be-tested SVG.

[0020] Preferably, in Step S23, the fitness function F in is defined as:

[0021]

[0022] In the formula, I and Q are the reactive current and reactive power output by the SVG identification model respectively; I * , Q * are the grid-connected voltage and reactive current output by the actual SVG controller respectively, obtained through testing on the RT-LAB platform; l is the total length of the data.

[0023] Preferably, Step S3 includes:

[0024] Step S31, initialize the parameters of the improved grey wolf algorithm according to the HIL data, and calculate the initial fitness with the objective function in S22;

[0025] Step S32, divide the wolf pack into α, β, and δ wolves according to the initial fitness, the optimal position is the α wolf, the sub-optimal position is the β wolf, and the third-optimal position is the δ wolf;

[0026] Step S33, the wolf pack enters the encirclement stage, and the mathematical expression is as follows:

[0027]

[0028] In the formula, t is the current iteration number; T is the maximum iteration number; A and C are two cooperative coefficient vectors; X p is the position vector of the prey; X is the position vector of the grey wolf; D represents the distance between the grey wolf and the prey; r1 and r2 are random numbers between 0 and 1; a is the convergence factor; it linearly decreases from 2 to 0 as the iteration number increases.

[0029] Step S34: Improve the mathematical expression in the wolf pack surrounding stage by introducing an improved convergence factor based on the sine function;

[0030] Step S35: Approach or move away from the prey position according to the value of |A|. When |A| < 1, the wolf pack will attack the prey, that is, make its search agent approach the prey position; when |A| > 1, the wolf pack will start to look for new prey, that is, make its search agent move away from the original prey position;

[0031] Step S36: The wolf pack enters the hunting stage, and the position update formula for each grey wolf is as follows:

[0032]

[0033] where D α , D β , D δ are the distances between the alpha, beta, and delta wolves and other individuals respectively; X α , X β , X δ are the current positions of the alpha, beta, and delta wolves respectively; C1, C2, and C3 are three random vectors; X is the position of the current wolf; t is the current iteration number;

[0034] Step S37: Improve the grey wolf position update formula in Step S36 by introducing a dynamic hunting weight coefficient;

[0035] Step S38: Randomly apply Logistic mapping perturbation;

[0036] Step S39: Calculate the error. If the error is satisfied, output the optimal parameters.

[0037] Preferably, in the said Step S34, the improved formula is as follows:

[0038]

[0039] Preferably, in the said Step S37, the improved position update formula is as follows:

[0040]

[0041] where ω α , ω β , ω δ are the hunting weight coefficients of the alpha, beta, and delta wolves at the t-th iteration respectively; ω α0 , ω β0 , ω δ0 are the initial hunting weight coefficients of the alpha, beta, and delta wolves respectively, satisfying ω α0 > ω β0 > ω δ0 ; τ α , τβ , τ δ are three attenuation constants respectively, satisfying τ α < τ β < τ δ .

[0042] Preferably, in the step S38, after the perturbation is applied, the position update formula becomes:

[0043]

[0044] In the formula, E is the magnitude of the currently applied perturbation; r is the mapping parameter, which is a constant on [3.57, 4] and takes 3.99; lg is a random number between 0 and 1, which plays a role in controlling the generation of the Logistic mapping perturbation, and the probability of setting the perturbation is 20%.

[0045] Preferably, in the step S39, the error is calculated by the following formula:

[0046]

[0047] In the formula, x mea and x sim are the measured data and the identified data respectively; δ ME is the average deviation of each interval; δ MAE is the average absolute deviation of each interval; δ G is the weighted average absolute deviation; δ MXE is the maximum deviation; K start and K end are the serial numbers of the starting and ending condition data respectively.

[0048] Advantages of the present invention: First, the control principle of SVG is analyzed and the electromechanical transient model of SVG is established in the present invention, and the parameters to be identified are clarified. Then, the HIL simulation system of the SVG controller is built to conduct HIL tests on SVG to obtain HIL data. After that, the original GWO algorithm is improved by using a fusion multi-strategy of dynamic hunting weight coefficient, non-linear improved convergence factor and randomly applying Logistic mapping perturbation, which greatly improves the convergence speed and accuracy of the algorithm. On this basis, the improved gray wolf algorithm is used to identify the SVG control parameters in combination with the HIL test data, which ensures both the identification accuracy and the identification efficiency. Using this method can achieve high-precision modeling of SVG to improve the correctness of the simulation of the power system with high new energy penetration rate, and provide a technical basis for the subsequent modeling of new energy power stations. Description of the Drawings

[0049] Figure 1 is the schematic diagram of the SVG control parameter identification method based on the improved gray wolf algorithm;

[0050] Figure 2 It is the electromechanical model diagram of the SVG controller;

[0051] Figure 3 Schematic diagram of the SVG controller RT-LAB hardware-in-the-loop simulation system;

[0052] Figure 4 Flowchart of the improved grey wolf algorithm;

[0053] Figure 5 It is the comparison chart of the iteration curves of the fitness functions of different algorithms;

[0054] Figure 6 Schematic diagram of the SVG fault interval division;

[0055] Figure 7 It is the comparison chart of the capacitive low voltage ride-through high power identification results;

[0056] Figure 8 It is the comparison chart of the inductive low voltage ride-through high power identification results. Specific implementation manners

[0057] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0058] Embodiment 1:

[0059] As Figure 1 , a method for identifying SVG control parameters based on an improved grey wolf algorithm, the steps include:

[0060] Step S1, analyze the control principle of the SVG and establish an SVG electromechanical transient model to clarify the parameters to be identified;

[0061] Step S2, perform hardware-in-loop (HIL) tests on the RT-LAB platform to obtain the data sets of the actual SVG controller under different fault conditions;

[0062] Step S3, based on the HIL data sets, use the improved grey wolf algorithm to identify the SVG control parameters and calculate their errors;

[0063] Further, the step S1 includes the following steps:

[0064] Step S11, as Figure 2 , establish an SVG electromechanical model according to the SVG control strategy in the PSASP power system analysis comprehensive program of the China Electric Power Research Institute.

[0065] Step S12, based on the control strategy in S11, the differential equation of the SVG controller is as follows:

[0066]

[0067] Wherein, V T1 is the output of the voltage measurement section, V T23 and V T45 are the outputs of the first and second lead-lag sections respectively, V p and V i are the outputs of the proportional section and the integral section respectively. U is the voltage at the output terminal of the SVG; U ref is the voltage reference value; U S is the AC side voltage; I S is the SVG output current; T1 is the time constant of the measurement loop; T2, T3, T4, T5 are four different lead-lag time constants in the PI regulator; T P is the proportional section time constant; T S is the trigger delay; K p , K i are the proportional and integral coefficients of the PI controller respectively; K D is the slope of the response characteristic curve; L is the filter reactance between the SVG and the system, which is a known quantity; U max , U min are the upper and lower limits of the voltage; I Cmax , I Lmax are the maximum capacitive and inductive currents; U smax , U smin are the limits of the proportional and integral sections

[0068] Step S13, based on the research in step S12, the parameter set to be identified by the SVG controller model is:

[0069] γ = [T1, T2, T3, T4, T5, K p , T p , K i ,

[0070] T S , K D , U max , U min , I Cmax , I Lmax

[0071] Furthermore, the specific steps of S2 are as follows;

[0072] Step S21, referring to Figure 3 , build a simulation system required for HIL testing. This system consists of a 220 kV infinite power supply, a transformer with a rated capacity of 100 MVA and a rated transformation voltage level of 220 / 35 kV, a filter inductor L and a to-be-tested SVG.

[0073] ​Step S22, HIL data acquisition: Based on the simulation system built in step S21, connect RT-LAB to the actual SVG controller through optical fiber, test the fault response characteristics of the SVG under various fault conditions, and generate corresponding simulation data sets.

[0074] Step S23, define the fitness function: The fitness function F in is defined as:

[0075]

[0076] In the formula, I and Q are the reactive current and reactive power output by the SVG identification model respectively; I * , Q * are the grid-connected voltage and reactive current output by the actual controller of the SVG respectively, obtained through testing on the RT-LAB platform; l is the total length of the data.

[0077] Furthermore, the specific steps of S3 are as follows:

[0078] Step S31, refer to Figure 4 , initialize the parameters of the improved grey wolf algorithm according to the HIL data, and calculate the initial fitness with the objective function in S22.

[0079] Step S32, divide the wolf pack into α, β and δ wolves according to the initial fitness. The optimal position is the α wolf, the second-best position is the β wolf, and the third-best position is the δ wolf.

[0080] Step S33, the wolf pack enters the encirclement stage, and the mathematical expression is as follows:

[0081]

[0082] In the formula, t is the current iteration number; T is the maximum iteration number; A and C are two cooperative coefficient vectors; X p is the position vector of the prey; X is the position vector of the grey wolf; D represents the distance between the grey wolf and the prey; r1 and r2 are random numbers between 0 and 1; a is the convergence factor; it linearly decreases from 2 to 0 as the iteration number increases.

[0083] Step S34, introduce an improved convergence factor based on the sine function to improve the mathematical expression of the wolf pack encirclement stage. The improved formula is as follows:

[0084]

[0085] Step S35, approach or move away from the prey position according to the value of |A|. When |A| < 1, the wolf pack will attack the prey, that is, make its search agent approach the prey position; when |A| > 1, the wolf pack starts to look for new prey, that is, make its search agent move away from the original prey position.

[0086] Step S36, the wolf pack enters the hunting stage, and the position update formula for each gray wolf is as follows:

[0087]

[0088] In the formula, D α , D β , D δ are the distances between the α, β, and δ wolves and other individuals respectively; X α , X β , X δ are the current positions of the α, β, and δ wolves respectively; C1, C2, and C3 are three random vectors; X is the position of the current wolf; and t is the current iteration number.

[0089] Step S37, introduce a dynamic hunting weight coefficient to improve the gray wolf position update formula in Step S36. The improved position update formula is as follows:

[0090]

[0091] In the formula, ω α , ω β , ω δ are the hunting weight coefficients of the α, β, and δ wolves at the t-th iteration respectively; ω α0 , ω β0 , ω δ0 are the initial hunting weight coefficients of the α, β, and δ wolves respectively, satisfying ω α0 > ω β0 > ω δ0 ; τ α , τ β , τ δ are three decay constants respectively, satisfying τ α < τ β < τ δ .

[0092] Step S38, randomly apply a Logistic mapping perturbation. After applying the perturbation, the position update formula becomes:

[0093]

[0094] In the formula, E is the magnitude of the currently applied perturbation; r is the mapping parameter, which is a constant in the range of [3.57, 4], taking 3.99; lg is a random number between 0 and 1, which plays a role in controlling the generation of the Logistic mapping perturbation, and the probability of setting the perturbation is 20%.

[0095] Step S39, referring to Figure 6 , calculate the error of each fault interval through the following formula. If the error is satisfied, output the optimal parameters.

[0096]

[0097] where x mea and x sim are the measured data and the identified data respectively; δ ME is the average deviation of each interval; δ MAE is the average absolute deviation of each interval; δ G is the weighted average absolute deviation; δ MXE is the maximum deviation; K start and K end are the sequence numbers of the starting and ending condition data respectively.

[0098] Example 2:

[0099] To verify the effectiveness of the proposed improved grey wolf algorithm, a hardware-in-the-loop simulation model is built based on the RT-LAB simulation platform. The HIL test system is as Figure 3 shown. Through the RT-LAB hardware-in-the-loop test platform, different fault conditions are simulated for semi-physical simulation, and a multi-condition HIL test data set is collected. The test conditions are shown in Table 1. Thereafter, based on the HIL data set, the improved grey wolf algorithm is used to identify the SVG control parameters and the identification results are brought into the model. Finally, the error between the simulation results and the test results is calculated to verify the effectiveness of the proposed identification method.

[0100] Table 1 Test Conditions

[0101]

[0102] To verify the effectiveness and superiority of the IGWO algorithm in the identification of SVG control output, the same SVG control measured data set is selected in this paper. The PSO, GA, GWO and IGWO algorithms are used for identification respectively, and the iteration curves of the fitness functions of different algorithms are shown in Figure 5 . The population size of all algorithms is 50, and the maximum number of iterations is set to 50 times.

[0103] As can be seen from Figure 5 , the PSO algorithm is prone to premature convergence during the optimization process and falls into the local optimum, resulting in poor algorithm accuracy; although the GA algorithm is not easily trapped in the local optimum trap through genetic and mutation methods, its convergence speed is slow and it cannot guarantee efficient application during the identification process; the original GWO algorithm shows a good convergence speed in the initial stage of convergence, but it is often prone to falling into the local optimum in the later stage.

[0104] Generally speaking, the IGWO algorithm has obvious advantages in terms of convergence speed and solution accuracy compared with other algorithms, verifying the feasibility of the IGWO algorithm.

[0105] Based on the HIL dataset, the improved grey wolf algorithm is used to identify the SVG control parameters. The identification results are brought into the PSASP electromechanical identification model, and the response curves of the grid-connected point voltage U, reactive current I q and reactive power Q output by the identification model are compared with the HIL measured results. The comparison results are shown in Figure 7 and Figure 8 . After that, combined with Figure 6 , the average identification error of each fault interval is calculated using the formula of S39. The error results are shown in Table 2, Table 3 and Table 4 to verify the feasibility and effectiveness of the proposed method.

[0106] Table 2 Calculation results of the error of the terminal voltage U

[0107]

[0108] Table 3 Calculation results of the error of the reactive current I

[0109]

[0110] Table 4 Calculation results of the reactive power Q

[0111]

[0112] It can be seen that the deviation under each working condition is less than 5%. Therefore, it can be obtained that the proposed SVG control parameter identification method based on the improved grey wolf algorithm has good identification accuracy and can meet the requirements of power system simulation.

[0113] The above embodiments are only the preferred technical solutions of the present invention and should not be regarded as limitations to the present invention. The protection scope of the present invention should be the technical solutions recorded in the claims, including equivalent replacement solutions of the technical features in the technical solutions recorded in the claims. That is, equivalent replacement improvements within this scope are also within the protection scope of the present invention.

Claims

1. An SVG control parameter identification method based on an improved grey wolf algorithm, characterized in that: It includes the following steps: Step S1: Analyze the control principle of SVG and establish an SVG electromechanical transient model to clarify the parameters to be identified; Step S2: Conduct hardware-in-the-loop tests on the RT-LAB platform to obtain the data sets of the actual SVG controller under different fault conditions; Step S3: Based on the HIL data sets, use the improved grey wolf algorithm to identify the SVG control parameters and calculate their errors.

2. The SVG control parameter identification method based on the improved grey wolf algorithm according to claim 1, wherein: The said Step S1 includes: Step S11: Establish an SVG electromechanical model according to the SVG control strategy in the PSASP power system analysis and synthesis program; Step S12: Based on the control strategy in S11, the differential equation of the SVG controller is as follows: Where, V T1 is the output of the voltage measurement section, V T23 and V T45 are the outputs of the first and second lead-lag sections respectively, V p and V i are the outputs of the proportional section and the integral section respectively; U is the voltage at the output terminal of the SVG; U ref is the voltage reference value; U S is the voltage on the AC side; I S is the SVG output current; T1 is the time constant of the measurement loop; T2, T3, T4, T5 are the four different lead-lag time constants in the PI regulator; T P is the proportional section time constant; T S is the trigger delay; K p , K i are the proportional and integral coefficients of the PI controller respectively; K D is the slope of the response characteristic curve; L is the filter reactance between the SVG and the system, which is a known quantity; U max , U min are the upper and lower limits of the voltage; I Cmax , I Lmax are the maximum capacitive and inductive currents; U smax , U smin are the limits of the proportional and integral sections; Step S13: Based on the research in Step S12, the parameter set to be identified for the SVG controller model is:

3. The SVG control parameter identification method based on the improved grey wolf algorithm according to claim 1, characterized in that: The said Step S2 includes: Step S21: Build the simulation system required for HIL tests; Step S22: HIL data acquisition: Based on the simulation system built in Step S21, connect the RT-LAB and the actual SVG controller through optical fibers, test the fault response characteristics of the SVG under various fault conditions, and generate the corresponding simulation data sets; Step S23: Define the fitness function.

4. The SVG control parameter identification method based on the improved grey wolf algorithm according to claim 1, characterized in that: In the said Step S21, the simulation system consists of a 220 kV infinite power source, a transformer with a rated capacity of 100 MVA and a rated voltage transformation level of 220 / 35 kV, a filtering inductor L, and an SVG to be tested.

5. The SVG control parameter identification method based on the improved grey wolf algorithm according to claim 1, wherein: In the step S23, the fitness function F in is defined as: Wherein, I and Q are the reactive current and reactive power output by the SVG identification model, respectively; I * , Q * are the grid-connected voltage and reactive current output by the actual controller of the SVG, respectively, obtained through testing on the RT-LAB platform; l is the total data length.

6. The SVG control parameter identification method based on the improved gray wolf algorithm according to claim 3, characterized in that: The said Step S3 includes: Step S31: Initialize the parameters of the improved grey wolf algorithm according to the HIL data, and calculate the initial fitness using the objective function in S22; Step S32: Divide the wolf pack into α, β, and δ wolves according to the initial fitness, with the optimal position being the α wolf, the second-best position being the β wolf, and the third-best position being the δ wolf; Step S33: The wolf pack enters the encirclement stage, and the mathematical expression is as follows: where t is the current iteration number; T is the maximum iteration number; A and C are two cooperative coefficient vectors; X p is the position vector of the prey; X is the position vector of the grey wolf; D represents the distance between the grey wolf and the prey; r1 and r2 are random numbers between 0 and 1; a is the convergence factor; linearly decreasing from 2 to 0 as the iteration number increases; Step S34: Introduce an improved convergence factor based on the sine function to improve the mathematical expression of the wolf pack encirclement stage; Step S35: Approach or move away from the prey position according to the value of |A|. When |A| < 1, the wolf pack will attack the prey, that is, make its search agent approach the prey position; when |A| > 1, the wolf pack starts to look for new prey, that is, make its search agent move away from the original prey position; Step S36: The wolf pack enters the hunting stage, and the position update formula for each grey wolf is as follows: where D α , D β , D δ are the distances between the alpha, beta, and delta wolves and other individuals respectively; X α , X β , X δ are the current positions of the alpha, beta, and delta wolves respectively; C1, C2, and C3 are three random vectors; X is the position of the current wolf; t is the current iteration number; Step S37: Introduce a dynamic hunting weight coefficient to improve the grey wolf position update formula in Step S36; Step S38: Randomly apply Logistic mapping perturbation; Step S39: Calculate the error. If the error is satisfied, output the optimal parameters.

7. The SVG control parameter identification method based on the improved grey wolf algorithm according to claim 6, characterized in that: In the said Step S34, the improved formula is as follows:

8. The SVG control parameter identification method based on the improved grey wolf algorithm according to claim 6, characterized in that: In the said Step S37, the improved position update formula is as follows: where ω α , ω β , ω δ are the hunting weight coefficients of the α, β, δ wolves at the t-th iteration, respectively; ω α0 , ω β0 , ω δ0 are the initial hunting weight coefficients of the α, β, δ wolves, respectively, satisfying ω α0 > ω β0 > ω δ0 ; τ α , τ β , τ δ are three decay constants, respectively, satisfying τ α < τ β < τ δ .

9. The SVG control parameter identification method based on the improved grey wolf algorithm according to claim 6, characterized in that: In the said Step S38, the position update formula after applying the perturbation becomes: Wherein, E is the magnitude of the currently applied perturbation; r is the mapping parameter, which is a constant on [3.57, 4], taking 3.99; lg is a random number between 0 and 1, which plays a role in controlling the generation of Logistic mapping perturbation, and the probability of setting the perturbation is 20%; 10. The SVG control parameter identification method based on the improved grey wolf algorithm according to claim 6, wherein: In the said Step S39, calculate the error through the following formula: where x mea and x sim are the measured data and the identified data respectively; δ ME is the average deviation of each interval; δ MAE is the average absolute deviation of each interval; δ G is the weighted average absolute deviation; δ MXE is the maximum deviation; K start and K end are the serial numbers of the starting and ending condition data respectively.

Images