Direct-driven fan parameter identification method based on improved sparrow algorithm

Abstract

The invention provides a direct-driven fan parameter identification method based on an improved sparrow algorithm, and relates to the technical field of wind power generation control, and the method comprises the steps: S1, determining a to-be-identified control parameter in a low-voltage ride-through control strategy according to an active power model and a reactive power model of a direct-driven fan; s2, identifying the to-be-identified control parameters of the direct-driven fan by adopting a sparrow algorithm; wherein in the step S2, the sparrow algorithm is improved by optimizing the search step sizes of the discoverer and the subscriber, the search step size of the predator and the initial position of the population and introducing a dynamic migration strategy. According to the scheme, the defects of optimization capacity and convergence speed can be improved, and the parameter identification efficiency is improved.

Description

Technical Field

[0001] The invention relates to the technical field of wind power generation control, and in particular to a direct-drive wind turbine parameter identification method based on an improved sparrow algorithm. Background Art

[0003] Due to its intermittent and random characteristics, wind power brings many uncertain challenges to the safe and stable operation of the power grid after being connected to the grid on a large scale through power electronic devices. In order to accurately analyze the impact of large-scale wind power on the power grid, it is necessary to establish an accurate wind turbine model, and the accuracy of the model depends to a large extent on the accuracy of the parameters. Unfortunately, in practice, the key parameters of wind turbines are usually not accurately known and are difficult to obtain directly from the nameplate or technical manual of the wind turbine. Therefore, in order to accurately analyze and study the impact of wind power on the power grid, it is necessary to identify the key parameters of the wind turbine.

[0004] The traditional Binary algorithm is known for its intuitive principle and convenient implementation, but its identification accuracy is often unsatisfactory when faced with multi-parameter identification tasks. Based on this, intelligent identification algorithms came into being, and intelligent identification algorithms cover multiple branches such as heuristic intelligent optimization algorithms and artificial intelligence algorithms. Heuristic intelligent optimization algorithms have shown significant advantages in the field of multi-parameter identification and can identify multiple parameters at the same time. However, the current intelligent identification algorithms based on heuristic intelligent optimization algorithms have poor optimization capabilities and convergence speed when performing parameter identification, require a lot of computing time, and have poor parameter identification efficiency. Summary of the invention

[0005] In view of this, in order to address the above shortcomings, it is necessary to propose a direct-drive wind turbine parameter identification method based on the improved sparrow algorithm to improve the defects of optimization ability and convergence speed and improve the efficiency of parameter identification.

[0006] The present invention provides a direct-drive fan parameter identification method based on an improved sparrow algorithm, comprising:

[0007] Step S1: determining the control parameters to be identified in the low voltage ride through control strategy according to the active power model and reactive power model of the direct drive wind turbine;

[0008] Step S2: using the sparrow algorithm to identify the control parameters to be identified of the direct-drive wind turbine;

[0009] Among them, in step S2, the sparrow algorithm is improved by optimizing the search step length of the discoverer and joiner, the search step length of the predator, the initial position of the population, and introducing a dynamic migration strategy.

[0010] Preferably, in step S1, the active power model of the direct-drive wind turbine is expressed as:

[0011] I pL =K Ip1 U t +K Ip2 I p0 +I pset

[0012] In the formula, I pL is the active current command, U t is the terminal voltage, K Ip1 and K Ip2 All are low-power active current calculation coefficients, I pset is the low active current setting value, I p0 is the initial active current;

[0013] The reactive power model of the direct-drive wind turbine is expressed as:

[0014] I qL =K Iq1 (U Lin -U t )+K Iq2 I q0 +I qset

[0015] In the formula, I qL is the reactive current command, U t is the terminal voltage, U Lin is the low voltage crossing threshold, K Iq1 and K Iq2 All are low-power reactive current calculation coefficients, I qset is the low reactive current setting value, I q0 is the initial reactive current;

[0016] K Ip1 , K Ip2 ,I pset , K Iq1 , K Iq2 ,I qset Determined as the control parameter to be identified in the low voltage ride through control strategy.

[0017] Preferably, the step S2 specifically includes:

[0018] Step S21: Each control parameter to be identified is represented by individual sparrows, and the group position is initialized using the following calculation formula:

[0019] X=l b +rand*(u b -l b )

[0020] In the formula, X is used to represent the group position, u b and lb They are used to characterize the upper and lower position boundaries of the population, respectively;

[0021] Step S22: Update the finder's location using the following calculation formula:

[0022]

[0023] In the formula, It is used to characterize the position of the i-th sparrow in the t+1th iteration, θ is a random constant between (0,1], i termax is the maximum number of iterations, R2 is the warning value in the range of (0,1], S safe is a safe value in the range of [0.5,1], Q is the search step size of the discoverer and joiner, and L is the unit column vector;

[0024] Step S23: Update the joiner's position using the following calculation formula:

[0025]

[0026] In the formula, is the global worst position of the t-th iteration, It is used to characterize the optimal position of the finder t+1 iteration, A + =A T (AA T ) -1 , A is a matrix whose elements are randomly assigned 1 or -1, and n is the number of sparrows;

[0027] Step S24: Randomly select a part of sparrows to perform the reconnaissance mission, as shown below:

[0028]

[0029] In the formula, represents the global optimal position of the t-th iteration, β is the search step of the predator, K is a uniform random number between [-1,1], and f i represents the fitness value of the i-th sparrow, f g represents the global optimal fitness value, f w Represents the global worst fitness value, and ε is a minimum value that prevents the denominator from being 0.

[0030] Preferably, in step S24, the fitness value is calculated by the following fitness function:

[0031]

[0032] In the formula, f(x i ) is used to represent the fitness value corresponding to the i-th sparrow individual, x iUsed to represent the i-th individual in the population, y true,k Used to represent the true value of the kth sample, y pred,k (x i ) is the use of individual x i The model represents the predicted value of the kth sample, where N is the number of samples.

[0033] Preferably, the optimization of the search step length of the finder and joiner includes:

[0034] An adaptive adjustment factor α is introduced. In each search, the search step length Q of the discoverer and joiner is optimized by the following calculation formula:

[0035]

[0036] Where Q′ is the optimized search step of the finder and joiner, H is the current search number, and H max The maximum number of searches.

[0037] Preferably, the optimization of the search step length of the predator includes:

[0038] Introducing the adjustment factors η and rand(-1,1), during each predator search, the predator search step length β is optimized by the following calculation formula:

[0039]

[0040] Where β′ is the optimized predator search step length, t is the current iteration number, and t max is the maximum number of iterations.

[0041] Preferably, the optimization of the initial position of the population includes:

[0042] A composite chaotic map is generated by fusing the Logistic chaotic map and the Tent chaotic map to initialize the initial position distribution of the sparrows.

[0043] Preferably, the composite chaotic mapping sequence is calculated by the following formula:

[0044]

[0045] In the formula, x i ' +1 is the composite chaotic map of the i+1th individual in the population, r is a control parameter with a value range of (0,4), and mod is a modulo operation.

[0046] Preferably, the introduction of a dynamic migration strategy to improve the sparrow algorithm includes:

[0047] After each iteration, a portion of individuals are randomly migrated using the following calculation formula:

[0048]

[0049] in, is the global worst position of the t-th iteration, represents the global optimal position of the t-th iteration, Used to represent the position of the i-th sparrow in the t+1 iteration.

[0050] Preferably, further comprising:

[0051] The disturbance process of the direct-drive fan is divided into three intervals: A, B, and C. The deviation of the parameter identification result is calculated using the following calculation formula to verify the accuracy of the parameter identification:

[0052]

[0053] In the formula, y obs and sim are the test value and model simulation value respectively, ε avg is the average deviation of the interval, ε abs is the mean absolute deviation, ε max is the maximum allowable value of the maximum deviation, ε G is the weighted mean absolute deviation, ε A , ε B , ε C are the mean absolute deviations of intervals A, B, and C, respectively, and N start and N final They are the serial numbers of the starting and ending working condition data respectively.

[0054] It can be seen from the above technical scheme that in the direct-drive wind turbine parameter identification method based on the improved sparrow algorithm provided by the present invention, it is considered to determine the control parameters to be identified in the low-voltage ride-through control strategy according to the active power model and reactive power model of the direct-drive wind turbine, and then the sparrow algorithm is used to identify the control parameters to be identified of the direct-drive wind turbine. Moreover, when the sparrow algorithm is used for parameter identification, it is considered to improve the sparrow algorithm by optimizing the search step size of the discoverer and the joiner, the search step size of the predator, the initial position of the population, and introducing a dynamic migration strategy. In this way, by optimizing the search step size and the initial position of the population, the defects of the traditional method in terms of optimization ability and convergence speed can be greatly improved, and by introducing a migration strategy, the algorithm is prevented from falling into a local optimal solution, which not only greatly saves the calculation time and improves the efficiency of parameter identification, but also makes the accuracy of parameter identification higher. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1A flow chart of a direct-drive wind turbine parameter identification method based on an improved sparrow algorithm provided in an embodiment of the present invention.

[0056] Figure 2 Schematic diagram of the five interval segments divided for the voltage ride-through process.

[0057] Figure 3 This is a schematic diagram of the low voltage ride through test.

[0058] Figure 4 Iteration graphs of error curves for SSA, ISSA and PSO.

[0059] Figure 5 P0 is 0.9pu, U t Comparison of simulation results of response curves of active power P and reactive power Q when the power consumption is 0.2pu.

[0060] Figure 6 P0 is 0.9pu, U t Comparison of simulation results of response curves of active power P and reactive power Q when the power consumption is 0.4 pu. DETAILED DESCRIPTION

[0061] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0062] like Figure 1 As shown, the present invention provides a direct-drive fan parameter identification method based on an improved sparrow algorithm, which may include the following steps:

[0063] Step S1: determining the control parameters to be identified in the low voltage ride through control strategy according to the active power model and reactive power model of the direct drive wind turbine;

[0064] Step S2: Use the sparrow algorithm to identify the control parameters to be identified of the direct-drive wind turbine; wherein, in step S2, the sparrow algorithm is improved by optimizing the search step length of the discoverer and joiner, the search step length of the predator, the initial position of the population, and introducing a dynamic migration strategy.

[0065] In this embodiment, it is considered to determine the control parameters to be identified in the low voltage ride through control strategy according to the active power model and reactive power model of the direct drive wind turbine, and then the sparrow algorithm is used to identify the control parameters to be identified of the direct drive wind turbine. Moreover, when the sparrow algorithm is used for parameter identification, it is considered to improve the sparrow algorithm by optimizing the search step size of the discoverer and the joiner, the search step size of the predator, the initial position of the population, and introducing a dynamic migration strategy. In this way, by optimizing the search step size and the initial position of the population, the defects of the traditional method in terms of optimization ability and convergence speed can be greatly improved, and by introducing a migration strategy, the algorithm is prevented from falling into a local optimal solution, which not only greatly saves the calculation time and improves the efficiency of parameter identification, but also makes the accuracy of parameter identification higher. In addition, compared with the existing intelligent identification method, this scheme does not need to pre-set the search range of the parameters to be identified to achieve accurate identification of multiple parameters at the same time, and has higher flexibility and practicality in actual application scenarios.

[0066] For step S1, the control parameters to be identified in the low voltage ride through control strategy are determined according to the active power model and reactive power model of the direct drive wind turbine;

[0067] Under normal operation of the power grid, the power factor of the wind turbine is constant. The positive direction is defined as the current direction flowing from the converter side to the grid side. At this time, the initial active current I flowing through the grid-connected inverter is p0 and initial reactive current I q0 They are:

[0068]

[0069] Where P0 is the initial active power per unit value of the wind turbine during normal operation, Q0 is the reactive power per unit value during normal operation, U0 is the initial voltage per unit value of the grid connection point, and the reference value is the rated voltage.

[0070] When a voltage drop occurs in the power grid, the dynamic reactive current increment of the wind turbine ΔI t Should meet:

[0071] ΔI t =K U (0.9-U t )I N

[0072] In the formula, I N is the rated current per unit value, U t K is the per unit value of the grid voltage drop depth, U is the dynamic reactive current support coefficient.

[0073] In this way, in the active power current control mode, the system active current command I pLThe calculation formula is:

[0074] I pL =K Ip1 U t +K Ip2 I p0 +I pset

[0075] In the formula, I pL is the active current command, U t is the terminal voltage, K Ip1 and K Ip2 All are low-power active current calculation coefficients, I pset is the low active current setting value, I p0 is the initial active current;

[0076] In the current control mode of reactive power, the low voltage crossing threshold U is introduced Lin , at this time the reactive current command I qL for:

[0077] I qL =K Iq1 (U Lin -U t )+K Iq2 I q0 +I qset

[0078] In the formula, I qL is the reactive current command, U t is the terminal voltage, U Lin is the low voltage crossing threshold, K Iq1 and K Iq2 All are low-power reactive current calculation coefficients, I qset is the low reactive current setting value, I q0 is the initial reactive current;

[0079] Then, in the current control mode of the low voltage ride through control strategy, K Ip1 , K Ip2 ,I pset , K Iq1 , K Iq2 ,I qset Determined as the control parameter to be identified in the low voltage ride through control strategy.

[0080] The voltage drop U at each depth is simulated by the low voltage generator t Electromagnetic transient simulation is performed, and the machine-end response curve obtained by the electromagnetic transient simulation is used as a reference result; the electrical quantities included in the machine-end response curve are the active power P and reactive power Q at the output end of the wind turbine.

[0081] In step S2, the sparrow algorithm is used to identify the control parameters to be identified of the direct-drive wind turbine;

[0082] This step can be specifically implemented through the following steps S21 to S24:

[0083] Step S21: Each control parameter to be identified is represented by individual sparrows, and the group position is initialized using the following calculation formula:

[0084] X=l b +rand*(u b -l b )

[0085] In the formula, X is used to represent the group position, u b and l b They are used to characterize the upper and lower position boundaries of the population, respectively;

[0086] In this step, further consideration is given to optimizing the initial position of the population. In the sparrow algorithm SSA, the position layout of the initial sparrow population in the solution space has a key influence on mining the optimal path. The standard SSA algorithm adopts a random distribution method for the position distribution of the initial sparrow population in the solution space, which is prone to uneven distribution, resulting in a decrease in the convergence speed. If the initial population is unevenly distributed in the solution space, the possibility of the algorithm finding the best path will be significantly prompted. The chaotic search algorithm has the advantages of randomness, comprehensive search and non-repetitiveness, so that in terms of exploring the solution space, due to the random search method of the standard SSA algorithm, the use of chaotic search can more effectively improve the speed and accuracy of the algorithm in finding the optimal solution. Therefore, it is considered to introduce a fusion of Logistic chaotic mapping and Tent chaotic mapping to generate a Logistic-Tent composite chaotic mapping to initialize the position distribution of the sparrow population. The chaotic system combines the complex dynamic characteristics of the Logistic system with the characteristics of fast iteration speed and strong autocorrelation of the Tent chaotic system, so that the system can generate more and more uniform random sequences than a single chaotic system in a short time. Specifically, the composite chaotic mapping sequence can be calculated by the following calculation formula:

[0087]

[0088] In the formula, x i ' +1 is the composite chaotic map of the i+1th individual in the population, r is a control parameter with a value range of (0,4), and mod is a modulo operation. Here, by taking the modulo 1, we ensure that the x value remains between 0 and 1.

[0089] Step S22: Update the finder's location using the following calculation formula:

[0090]

[0091] In the formula, It is used to characterize the spatial position of the i-th sparrow in the t+1th iteration. The search space is d-dimensional, θ is a random constant between (0,1], and i termax is the maximum number of iterations, R2 is the warning value in the range of (0,1], S safe is a safe value in the range of [0.5,1], Q is the search step of the discoverer and the joiner, which is a normally distributed random number, and L is a d-dimensional unit column vector. When the predation environment is safe, that is, R2 safe When the hunting environment is dangerous, an alarm will be sounded, indicating that the discoverer should go to other locations to hunt.

[0092] Step S23: Update the joiner's position using the following calculation formula:

[0093]

[0094] In the formula, is the global worst position of the t-th iteration, It is used to characterize the optimal position of the finder t+1 iteration, A + =A T (AA T ) -1 , A is a 1*d matrix with randomly assigned elements of 1 or -1, and n is the number of sparrows; when i>n / 2, it means that the i-th participant has not obtained food and is in a hungry state. At this time, he needs to fly to other places to forage for more energy.

[0095] For steps S22-S23, since the search step length Q of the discoverer and joiner in the SSA algorithm is fixed, it does not have the ability to self-regulate. As the number of searches increases, the search step length should be more refined to explore potential optimization directions. Therefore, consider introducing an adaptive adjustment factor α, and optimize the search step length Q of the discoverer and joiner through the following calculation formula in each search:

[0096]

[0097] Where Q′ is the optimized search step of the finder and joiner, H is the current search number, and H max is the maximum number of searches, 0<α<1.

[0098] Step S24: Randomly select a part of sparrows to perform the reconnaissance mission, as shown below:

[0099]

[0100] In the formula,​ represents the global optimal position of the t-th iteration, β is the search step of the predator, is a random number from the standard normal distribution, K is a uniform random number in [-1,1], and f i represents the fitness value of the i-th sparrow, f g represents the global optimal fitness value, f w Represents the global worst fitness value, and ε is a minimum value that prevents the denominator from being 0.

[0101] In this step, since the predator's siege step length should be gradually reduced when it is considered to be closer to the optimal solution in the predation behavior, we consider introducing the adjustment factor η and rand(-1,1), and optimize the predator's search step length β by the following calculation formula every time the predator searches:

[0102]

[0103] Where β′ is the optimized predator search step length, t is the current iteration number, and t max is the maximum number of iterations, 0<η<1.

[0104] In addition, in order to prevent the algorithm from falling into a local optimal solution, it is also considered to randomly migrate some individuals after each iteration through the following calculation formula:

[0105]

[0106] in, is the global worst position of the t-th iteration, represents the global optimal position of the t-th iteration, Used to represent the position of the i-th sparrow in the t+1 iteration.

[0107] In order to verify the accuracy of the modeling and identification results, it is considered to verify the accuracy by calculating the deviation between the model simulation and the test results. Figure 2 As shown, the disturbance process is divided into five intervals: A, B1, B2, C1, and C2. The whole process can be composed of a steady-state interval and a transient interval. The deviation of the parameter identification result is calculated using the following formula to verify the accuracy of the parameter identification:

[0108]

[0109] In the formula, y obs and sim are the test value and model simulation value respectively, ε avg is the average deviation of the interval, ε abs is the mean absolute deviation, ε max is the maximum allowable value of the maximum deviation, ε G is the weighted mean absolute deviation, εA , ε B , ε C are the mean absolute deviations of intervals A, B, and C, respectively, and N start and N final They are the serial numbers of the starting and ending working condition data respectively.

[0110] The effect of this solution is described below with specific examples.

[0111] In order to verify the effectiveness of this solution, an electromagnetic transient simulation model is built based on the PLECS simulation platform. Figure 3 The schematic diagram of the low voltage ride-through test is shown in the figure. The low voltage generator simulates voltage drops of different depths to perform electromagnetic transient simulation, and the machine-side response curve obtained by the electromagnetic transient simulation is used as the test result. The test conditions are shown in Table 1, and then the identification results are brought into the model. Finally, the simulation results are compared with the test results to verify the effectiveness of the method represented.

[0112] Table 1

[0113]

[0114] In order to further verify the effect of this scheme, we consider comparing it with other algorithms, focusing on the two aspects of calculation speed and calculation accuracy. Active power and reactive power are selected as observations to identify the parameters, and the identification results of the improved sparrow algorithm ISSA are compared with those of the particle swarm algorithm PSO and the ordinary sparrow algorithm SSA.

[0115] First, the number of iterations is selected as 60, and 60 simulations are performed. The average value is taken as the result of parameter identification, and the objective function convergence curves of the three different optimization algorithms when solving the control parameters are obtained, as shown in Figure 4 shown.

[0116] In the simulation, it can be seen that when the maximum number of iterations is 50, the iteration speed of the improved sparrow algorithm exceeds that of the ordinary sparrow algorithm and the traditional particle swarm algorithm, and has a great advantage over both.

[0117] The identification results are brought into different working conditions, the errors between simulation and test values ​​are calculated, and the accuracy of modeling is verified by calculating the deviation between model simulation and test results P and Q. The simulation results detect the response curves of active power P and reactive power Q, such as Figure 5 and Figure 6 The figure shows the comparison between the simulation results and the test results. It can be seen from the figure that the accuracy of the identification results is high.

[0118] Furthermore, the simulation model results are compared with the maximum allowable values ​​specified in the standard as shown in Table 2-4:

[0119] Table 2 Maximum allowable deviation values ​​specified in the electromechanical transient model verification standard

[0120]

[0121] Table 3 Simulation model deviation calculation table (working condition 1)

[0122]

[0123] Table 4 Simulation model deviation calculation table (working condition 2)

[0124]

[0125] It can be seen from the above table that the deviations under each working condition are all less than the maximum allowable deviation value specified in Table 2. That is, the simulation results show that the identification results also have good adaptability to the working conditions in the nonlinear working area.

[0126] The present specification also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed in a computer, the computer is caused to execute a method in any one of the embodiments in the specification.

[0127] The present specification also provides a computing device, including a memory and a processor, wherein executable codes are stored in the memory, and when the processor executes the executable codes, a method in any embodiment of the present specification is implemented.

[0128] Since the device embodiments provided by the present invention are based on the same inventive concept as the method embodiments of this specification, the specific contents can be found in the description of the method embodiments of this specification and will not be repeated here.

[0129] The modules or units in the device of the embodiment of the present invention can be combined, divided and deleted according to actual needs. The above disclosure is only the preferred embodiment of the present invention, and of course it cannot be used to limit the scope of the rights of the present invention. Those skilled in the art can understand that all or part of the processes of the above embodiment are implemented, and the equivalent changes made according to the claims of the present invention still fall within the scope of the invention.

Claims

1. A direct-drive fan parameter identification method based on an improved sparrow algorithm, characterized in that: include: Step S1: determining the control parameters to be identified in the low voltage ride through control strategy according to the active power model and reactive power model of the direct drive wind turbine; Step S2: using the sparrow algorithm to identify the control parameters to be identified of the direct-drive wind turbine; Among them, in step S2, the sparrow algorithm is improved by optimizing the search step length of the discoverer and joiner, the search step length of the predator, the initial position of the population, and introducing a dynamic migration strategy.

2. The direct-drive wind turbine parameter identification method based on the improved sparrow algorithm according to claim 1 is characterized in that: In step S1, the active power model of the direct-drive wind turbine is expressed as: I pL =K Ip1 U t +K Ip2 I p0 +I pset In the formula, I pL is the active current command, U t is the terminal voltage, K Ip1 and K Ip2 All are low-power active current calculation coefficients, I pset is the low active current setting value, I p0 is the initial active current; The reactive power model of the direct-drive wind turbine is expressed as: AND qL =K Iq1 (IN Lin -IN t )+K Iq2 AND q0 +I qset In the formula, I qL is the reactive current command, U t is the terminal voltage, U Lin is the low voltage crossing threshold, K Iq1 and K Iq2 All are low-power reactive current calculation coefficients, I qset is the low-pass reactive current setting value, I q0 is the initial reactive current; K Ip1 , K Ip2 ,I pset , K Iq1 , K Iq2 ,I qset Determined as the control parameter to be identified in the low voltage ride through control strategy.

3. The direct-drive wind turbine parameter identification method based on the improved sparrow algorithm according to claim 1 is characterized in that: The step S2 specifically includes: Step S21: Each control parameter to be identified is represented by individual sparrows, and the group position is initialized using the following calculation formula: X=l b +rand*(u b -l b ) In the formula, X is used to represent the group position, u b and l b They are used to characterize the upper and lower position boundaries of the population, respectively; Step S22: Update the finder's location using the following calculation formula: In the formula, It is used to characterize the position of the i-th sparrow in the t+1th iteration, θ is a random constant between (0,1], i termax is the maximum number of iterations, R2 is the warning value in the range of (0,1], S safe is a safe value in the range of [0.5,1], Q is the search step size of the discoverer and joiner, and L is the unit column vector; Step S23: Update the joiner's position using the following calculation formula: In the formula, is the global worst position of the t-th iteration, It is used to characterize the optimal position of the finder t+1 iteration, A + =A T (AA T ) -1 , A is a matrix whose elements are randomly assigned 1 or -1, and n is the number of sparrows; Step S24: Randomly select a part of sparrows to perform the reconnaissance mission, as shown below: In the formula, represents the global optimal position of the t-th iteration, β is the search step of the predator, K is a uniform random number between [-1,1], and f i represents the fitness value of the i-th sparrow, f g represents the global optimal fitness value, f w Represents the global worst fitness value, and ε is a minimum value that prevents the denominator from being 0.

4. The direct-drive wind turbine parameter identification method based on the improved sparrow algorithm according to claim 3 is characterized in that: In step S24, the fitness value is calculated by the following fitness function: In the formula, f(x i ) is used to represent the fitness value corresponding to the i-th sparrow individual, x i Used to represent the i-th individual in the population, y true,k Used to represent the true value of the kth sample, y pred,k (x i ) is the use of individual x i The model represents the predicted value of the kth sample, where N is the number of samples.

5. The direct-drive wind turbine parameter identification method based on the improved sparrow algorithm according to claim 3 is characterized in that: The optimization of the search step length of the finder and joiner includes: An adaptive adjustment factor α is introduced. In each search, the search step length Q of the discoverer and joiner is optimized by the following calculation formula: Where Q′ is the optimized search step of the finder and joiner, H is the current search number, and H max The maximum number of searches.

6. The direct-drive wind turbine parameter identification method based on the improved sparrow algorithm according to claim 3 is characterized in that: The optimization of the search step length of the predator includes: Introducing the adjustment factors η and rand(-1,1), during each predator search, the predator search step length β is optimized by the following calculation formula: Where β′ is the optimized predator search step length, t is the current iteration number, and t max is the maximum number of iterations.

7. The direct-drive wind turbine parameter identification method based on the improved sparrow algorithm according to claim 3 is characterized in that: The optimization of the initial position of the population includes: A composite chaotic map is generated by fusing the Logistic chaotic map and the Tent chaotic map to initialize the initial position distribution of the sparrows.

8. The direct-drive wind turbine parameter identification method based on the improved sparrow algorithm according to claim 7 is characterized in that: The composite chaotic mapping sequence is calculated by the following formula: In the formula, x i ' +1 is the composite chaotic map of the i+1th individual in the population, r is a control parameter with a value range of (0,4), and mod is a modulo operation.

9. The direct-drive wind turbine parameter identification method based on the improved sparrow algorithm according to claim 3 is characterized in that: The dynamic migration strategy is introduced to improve the sparrow algorithm, including: After each iteration, a portion of individuals are randomly migrated using the following calculation formula: in, is the global worst position of the t-th iteration, represents the global optimal position of the t-th iteration, Used to represent the position of the i-th sparrow in the t+1 iteration.

10. The direct-drive wind turbine parameter identification method based on the improved sparrow algorithm according to claim 1, characterized in that: Further including: The disturbance process of the direct-drive fan is divided into three intervals: A, B, and C. The deviation of the parameter identification result is calculated using the following calculation formula to verify the accuracy of the parameter identification: In the formula, y obs and sim are the test value and model simulation value respectively, ε avg is the average deviation of the interval, ε abs is the mean absolute deviation, ε max is the maximum allowable value of the maximum deviation, ε G is the weighted mean absolute deviation, ε A , ε B , ε C are the mean absolute deviations of intervals A, B, and C, respectively, and N start and N final They are the serial numbers of the starting and ending working condition data respectively.

Images