Direct-driven fan low-voltage ride-through control parameter identification method based on improved bald eagle search algorithm

Abstract

A direct-driven fan low voltage ride through (LVRT) control parameter identification method based on an improved bald eagle search algorithm comprises the following steps that S1, an electromechanical transient model of a direct-driven fan is built according to physical characteristics of the direct-driven fan, and LVRT control parameters, needing to be identified, of a control mode under all control strategies are determined; s2, collecting transient operation data of the direct-driven fan under various LVRT working conditions, processing the transient operation data, and taking the processed transient operation data as a test data set; and S3, based on the test data set obtained in the S1, identifying the control parameters of the direct-drive fan LVRT by adopting an improved Balde Eagle Search Algorithm (IBES).

Description

Technical Field

[0001] The present invention relates to the technical field of wind power generation control, and in particular to a method for identifying low voltage ride-through control parameters of a direct-drive wind turbine based on an improved Bald Eagle Search Algorithm (IBES). Background Art

[0002] However, due to the volatility and randomness of wind resources, the wind turbine output is uncertain. Therefore, the direct grid connection of wind turbines will inevitably bring many uncertain challenges to the safe and stable operation of the power grid. In order to improve the ability of the power system to operate safely and stably, it is necessary to conduct rapid and accurate simulation analysis of the large power grid, among which the model accuracy of the unit is particularly critical. The premise of establishing a high-precision model is to obtain the control parameters inside its controller. At present, the electromechanical transient simulation model is the most widely used. In the electromechanical transient simulation model of the wind turbine, the electrical characteristics during the low voltage ride through (LVRT, LowVotalge Ride Through) period are currently considered. Therefore, it is very important to obtain accurate LVRT control parameters. Due to the large number of wind power equipment manufacturers in my country, the equipment structures of different manufacturers and models are different, the grid connection characteristics are significantly different, and due to manufacturer confidentiality reasons, the accuracy of LVRT control parameters cannot be ensured. Therefore, LVRT control parameter identification based on test data is an effective way to obtain the control parameters of the new energy power generation system. However, when using traditional algorithms to identify LVRT control parameters, on the one hand, the influence of outliers is not taken into account, which causes outliers to interfere with the identification process, so that the traditional algorithm can only identify a single working condition, but has poor adaptability to the identification of multiple working conditions; on the other hand, due to the poor global search ability of the traditional algorithm, it is easy to fall into the local optimal solution, resulting in low identification accuracy and efficiency. Summary of the invention

[0003] In view of this, it is necessary to provide a low voltage ride through control parameter identification method for direct-drive wind turbines based on an improved vulture search algorithm to improve the identification accuracy and efficiency, and enhance the adaptability to multiple working conditions.

[0004] A method for identifying low voltage ride-through control parameters of a direct-drive wind turbine based on an improved vulture search algorithm comprises the following steps:

[0005] S1. Build the electromechanical transient model of the direct-drive wind turbine according to its physical characteristics, and determine the LVRT control parameters of the direct-drive wind turbine that need to be identified in the control mode under each control strategy;

[0006] S2. Collect transient operation data of the direct-drive wind turbine under various LVRT conditions, and process the transient operation data as a test data set;

[0007] S3. Based on the test data set obtained in S1, the improved vulture search algorithm IBES is used to identify the LVRT control parameters of the direct-drive wind turbine.

[0008] Preferably, step S1 specifically comprises the following steps:

[0009] S10. According to the physical characteristics of the direct-drive wind turbine, an electromechanical transient model of the direct-drive wind turbine is constructed, the dynamic response process of the inner loop control and the phase-locked loop of the full power converter is ignored, and the operating characteristics of the LVRT process are considered;

[0010] S11. According to the active power control strategy and reactive power control strategy of the direct-drive wind turbine converter, the specified current control mode under the two control strategies is analyzed:

[0011] Among them, the specified current control mode under the active power control strategy is:

[0012] Active current command I during LVRT pLVRT , obtained by formula (1), where K 1,Ip,LV and K 2,Ip,Lv All are low-power active current calculation coefficients; I pset It is the low active current setting value; I p0 is the initial active current; U t is the voltage drop degree;

[0013] I pLVRT =K 1,Ip,LV U t +K 2,Ip,LV I p0 +I pset Formula (1)

[0014] Specified current control mode under reactive power control strategy:

[0015] Reactive current command I during LVRT qLVRT Obtained by formula (2), where K 1,Iq,LV and K 2,Iq,LV All are low-power reactive current calculation coefficients; U Lin is the low penetration threshold; I qset It is the setting value of low reactive current; I q0 is the initial reactive current;

[0016] I qLVRT =K 1,Iq,LV (U Lin -U t )+K 2,Iq,LV I q0 +I qset Formula (2)

[0017] S12. According to the control logic of the specified current control mode under the two control strategies of the direct-drive wind turbine converter, determine the direct-drive wind turbine LVRT control parameters that need to be identified, including: low-pass active current calculation coefficient K 1,Ip,LV and K 2,Ip,Lv , low wear active current setting value I pset , low-power reactive current calculation coefficient K 1,Iq,LV and K 2,Iq,LV , low reactive current setting value I qset .

[0018] Preferably, step S3 specifically includes the following steps:

[0019] S30, based on the test data set, obtaining the initial value of the LVRT control parameter to be identified, and using the initial value of the LVRT control parameter as the initial data for IBES operation to determine the optimization range of the LVRT control parameter to be identified;

[0020] S31. Construct the objective function J with the root mean square error, as shown in formula (3), where N is the number of test data sets, I p,e ,I q,e , P e , Q e are the deviations between the corresponding active current, reactive current, active power, reactive power data in the test data set and the identification data; (substitute the initial values ​​of the control parameters into the model of step S1 to obtain the identification data)

[0021]

[0022] S32, initializing basic iteration parameters, including population size, number of iterations and spatial dimension;

[0023] S33. Use Bernoulli chaotic mapping to initialize the vulture population. The Bernoulli chaotic mapping is shown in formula (4), where μ1 is the chaos parameter with a value of 0.4, and y k is the current chaotic variable value, y k+1 is the updated chaotic variable value;

[0024]

[0025] S34, using formula (3) to calculate the fitness value of each vulture in the population, find the current optimal fitness value, that is, the minimum fitness value, and the corresponding position, set the search space with the position corresponding to the current optimal fitness value as the center, and update the position of each vulture according to the position corresponding to the current optimal fitness value, so that it moves toward the position corresponding to the optimal fitness value. The position of each vulture represents a set of LVRT control parameters. The position update formula is shown in formula (5), where Pi , P new,i are the original position and updated position of the i-th vulture, i is a parameter that controls the position change, ranging from 1 to 2, r is a random value between 0 and 1, and P best is the position of the vulture with the best fitness value in the population, P mean is the position mean;

[0026] P new,i =P best +sr(P mean -P i ) Formula (5)

[0027] S35. From the individuals with better fitness values, select the position of the individuals as the search space. In the set search space, the vulture spirals in different directions. The update formulas of the spiral flight position are shown in formulas (6), (7) and (8), where P i , P new,i are the original position and updated position of the i-th vulture, respectively, P i+1 is the original position of the i+1th vulture; α is used to determine the search angle between the vulture and the center point, R rand It is used to determine the number of search cycles, and its value range is [0,1]; θ(i) and r(i) are the polar angle and polar diameter of each vulture's spiral flight, respectively, and y(i) and x(i) are the positions of the vulture in polar coordinates;

[0028] P new,i =P i +y(i)(P i -P i+1 )+x(i)(P i -P mean ) Formula (6)

[0029]

[0030] S36, the vulture swings from the current best position in the search space to the global best position, that is, for each vulture, according to the result of the fitness value, compare the advantages and disadvantages of its individual best position and the global best position; if the current position is better, use formulas (9) and (10) to update the individual best position; if there is a better position in the vulture population, then this better position is updated to the global best position; where c1 and c2 represent the vulture's movement towards the center point, that is, the position mean P mean and the global optimal position P mean The strength of the movement, the value range is [1,2]; ω is the adaptive weight; t is the number of iterations of the algorithm this time;

[0031] P new,i =R rand Pbest +ω[x(i)(P i -c1P mean )+y(i)(P i -c2P best )] Formula (9)

[0032]

[0033] S37, repeating steps S35 and S36 until a predetermined number of iterations or an objective function is reached, that is, the fitness value is less than a given convergence error;

[0034] S38, obtaining the final global optimal position, and extracting the direct-drive wind turbine LVRT control parameters that need to be identified therefrom.

[0035] Preferably, in step S30, the least squares method is used to fit the initial value of the LVRT control parameter to be identified.

[0036] Preferably, step S2 specifically includes the following steps:

[0037] S20, set up multiple LVRT fault conditions for direct-drive fans, including high power, medium power and low power conditions;

[0038] S21, collecting transient operation data of the direct-drive wind turbine under the above three LVRT conditions;

[0039] S22, deleting duplicate records and samples containing missing values ​​in the transient operation data to obtain first intermediate data;

[0040] S23, calculating the mean μ2 and standard deviation σ2 of the first intermediate data, identifying data points that are beyond the range of the mean μ2 plus or minus three times the standard deviation σ2 according to the 3σ criterion, and removing these data points from the first intermediate data to obtain second intermediate data;

[0041] S24. Normalize the second intermediate data to obtain a test data set.

[0042] Preferably, in step S22, the mean μ2 and the standard deviation σ2 are obtained by formulas (11) and (12), respectively, where a i is the data sample of the first intermediate data, n is the total sample size of the first intermediate data; the minimum value in the range of the mean plus or minus three times the standard deviation is min3σ2, and the maximum value is max3σ2, which are obtained by formulas (13) and (14) respectively.

[0043]

[0044] min3σ2=μ2-3σ2 Formula (13)

[0045] max3σ2=μ2+3σ2 Formula (14)

[0046] Preferably, step S24 specifically includes the following steps:

[0047] S240, respectively obtain the mean μ3 and standard deviation σ3 of the second intermediate data by formula (15) and (16), where b i is a data sample of the second intermediate data, and n is the total number of samples of the second intermediate data;

[0048]

[0049] S241. Normalize the second intermediate data using formula (17) to obtain a test data set, where B i is the normalized data.

[0050]

[0051] The above-mentioned method for identifying low voltage ride-through control parameters of direct-drive wind turbines based on the improved vulture search algorithm, on the one hand, the test data set used is initially collected by the direct-drive wind turbines under various LVRT conditions, and then repeated values, missing values, abnormal values ​​are eliminated, and normalized. Compared with the prior art, the adaptability to multiple conditions is enhanced, and the accuracy of LVRT control parameter identification is improved; on the other hand, in the identification process, the least squares method is used to obtain preliminary results of LVRT control parameters, which narrows the optimization range and improves the identification efficiency. The Bernoulli chaotic mapping is further used to improve the initial global search capability of the algorithm, and finally the adaptive weight ω is used to increase the optimization strength of the early stage of the identification algorithm to avoid falling into the local optimum, while reducing the optimization range of the late stage of the identification algorithm, thereby improving the accuracy of identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is the overall flow chart of the present invention.

[0053] Figure 2 This is the process of identifying the LVRT control parameters of a direct-drive wind turbine using an improved vulture search algorithm in the present invention.

[0054] Figure 3 It is a comparison chart between the simulation results and the test results of working condition 1 in specific embodiment 1.

[0055] Figure 4 It is a comparison chart of the simulation results and the test results of working condition 2 in specific embodiment 1.

[0056] Figure 5 It is a comparison chart between the simulation results and the test results of working condition 3 in specific embodiment 1.

[0057] Figure 6 It is a comparison diagram of root mean square error iteration curves of the three algorithms in specific embodiment 2. DETAILED DESCRIPTION

[0058] The technical solutions and technical effects of the embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings of the present invention.

[0059] Please see Figure 1 As shown, a method for identifying low voltage ride-through control parameters of a direct-drive wind turbine based on an improved vulture search algorithm comprises the following steps:

[0060] S1. Build the electromechanical transient model of the direct-drive wind turbine according to its physical characteristics, and determine the LVRT control parameters of the direct-drive wind turbine that need to be identified in the control mode under each control strategy;

[0061] S2. Collect transient operation data of the direct-drive wind turbine under various LVRT conditions, and process the transient operation data as a test data set;

[0062] S3. Based on the test data set obtained in S1, the improved vulture search algorithm IBES is used to identify the LVRT control parameters of the direct-drive wind turbine.

[0063] Furthermore, step S1 specifically includes the following steps:

[0064] S10. According to the physical characteristics of the direct-drive wind turbine, an electromechanical transient model of the direct-drive wind turbine is constructed, the dynamic response process of the inner loop control and the phase-locked loop of the full power converter is ignored, and the operating characteristics of the LVRT process are considered;

[0065] S11. According to the active power control strategy and reactive power control strategy of the direct-drive wind turbine converter, the specified current control mode under the two control strategies is analyzed:

[0066] Among them, the specified current control mode under the active power control strategy is:

[0067] Active current command I during LVRT pLVRT , obtained by formula (1), where K 1,Ip,LV and K 2,Ip,Lv All are low-power active current calculation coefficients; I pset It is the low active current setting value; I p0 is the initial active current; U t is the voltage drop degree;

[0068] I pLVRT =K 1,Ip,LV U t +K 2,Ip,LV I p0 +Ipset Formula (1)

[0069] Specified current control mode under reactive power control strategy:

[0070] Reactive current command I during LVRT qLVRT Obtained by formula (2), where K 1,Iq,LV and K 2,Iq,LV All are low-power reactive current calculation coefficients; U Lin is the low penetration threshold; I qset It is the setting value of low reactive current; I q0 is the initial reactive current;

[0071] I qLVRT =K 1,Iq,LV (U Lin -U t )+K 2,Iq,LV I q0 +I qset Formula (2)

[0072] S12. According to the control logic of the specified current control mode under the two control strategies of the direct-drive wind turbine converter, determine the direct-drive wind turbine LVRT control parameters that need to be identified, including: low-pass active current calculation coefficient K 1,Ip,LV and K 2,Ip,Lv , low wear active current setting value I pset , low-power reactive current calculation coefficient K 1,Iq,LV and K 2,Iq,LV , low reactive current setting value I qset .

[0073] For further information, see Figure 2 As shown, step S3 specifically includes the following steps:

[0074] S30, based on the test data set, obtaining the initial value of the LVRT control parameter to be identified, and using the initial value of the LVRT control parameter as the initial data for IBES operation to determine the optimization range of the LVRT control parameter to be identified;

[0075] S31. Construct the objective function J with the root mean square error, as shown in formula (3), where N is the number of test data sets, I p,e ,I q,e , P e , Q e are the deviations between the corresponding active current, reactive current, active power, reactive power data in the test data set and the identification data;

[0076]

[0077] In this setting, identification data is obtained by substituting the initial values ​​of the control parameters into the model in step S1;

[0078] S32, initializing basic iteration parameters, including population size, number of iterations and spatial dimension;

[0079] S33. Use Bernoulli chaotic mapping to initialize the vulture population. The Bernoulli chaotic mapping is shown in formula (4), where μ1 is the chaos parameter with a value of 0.4, and y k is the current chaotic variable value, y k+1 is the updated chaotic variable value;

[0080]

[0081] Initializing the vulture population through chaotic mapping can affect the initial position and diversity of the population, improve the uniformity of the population's spatial distribution, and improve the initial global search capability of the algorithm;

[0082] S34, using formula (3) to calculate the fitness value of each vulture in the population, find the current optimal fitness value, that is, the minimum fitness value, and the corresponding position, set the search space with the position corresponding to the current optimal fitness value as the center, and update the position of each vulture according to the position corresponding to the current optimal fitness value, so that it moves toward the position corresponding to the optimal fitness value. The position of each vulture represents a set of LVRT control parameters. The position update formula is shown in formula (5), where P i , P new,i are the original position and updated position of the i-th vulture, i is a parameter that controls the position change, ranging from 1 to 2, r is a random value between 0 and 1, and P best is the position of the vulture with the best fitness value in the population, P mean is the position mean;

[0083] P new,i =P best +sr(P mean -P i ) Formula (5)

[0084] S35. From the individuals with better fitness values, select the position of the individuals as the search space. In the set search space, the vulture spirals in different directions. The update formulas of the spiral flight position are shown in formulas (6), (7) and (8), where P i , P new,i are the original position and updated position of the i-th vulture, respectively, P i+1 is the original position of the i+1th vulture; α is used to determine the search angle between the vulture and the center point, R randIt is used to determine the number of search cycles, and its value range is [0,1]; θ(i) and r(i) are the polar angle and polar diameter of each vulture's spiral flight, respectively, and y(i) and x(i) are the positions of the vulture in polar coordinates;

[0085] P new,i =P i +y(i)(P i -P i+1 )+x(i)(P i -P mean ) Formula (6)

[0086]

[0087] In this step, by making the vulture fly in spirals to search in different directions, the search range can be expanded to avoid falling into the local optimal solution;

[0088] S36, the vulture swings from the current best position in the search space to the global best position, that is, for each vulture, according to the result of the fitness value, compare the advantages and disadvantages of its individual best position and the global best position; if the current position is better, use formulas (9) and (10) to update the individual best position; if there is a better position in the vulture population, then this better position is updated to the global best position; where c1 and c2 represent the vulture's movement towards the center point, that is, the position mean P mean and the global optimal position P mean The strength of the movement, the value range is [1,2]; ω is the adaptive weight; t is the number of iterations of the algorithm this time;

[0089] P new,i =R rand P best +ω[x(i)(P i -c1P mean )+y(i)(P i -c2P best )] Formula (9)

[0090]

[0091] In this step, t has a larger value in the early stage of the iteration, which enables better global search, and decreases adaptively in the later stage of the iteration, which enables better local search and improves the convergence speed;

[0092] S37, repeating steps S35 and S36 until a predetermined number of iterations or an objective function is reached, that is, the fitness value is less than a given convergence error;

[0093] S38, obtaining the final global optimal position, and extracting the direct-drive wind turbine LVRT control parameters that need to be identified therefrom.

[0094] Furthermore, in step S30, the least squares method is used to fit the initial value of the LVRT control parameter to be identified.

[0095] Furthermore, step S2 specifically includes the following steps:

[0096] S20, set up multiple LVRT fault conditions for direct-drive fans, including high power, medium power and low power conditions;

[0097] S21, collecting transient operation data of the direct-drive wind turbine under the above three LVRT conditions; since it is difficult to obtain data from actual operation, the transient operation data in this step is obtained based on the electromagnetic transient model of the direct-drive wind turbine. The electromagnetic transient model is an equivalent model built according to the actual operation state. The data obtained from the electromagnetic transient model is equivalent to the data obtained from the actual operation;

[0098] S22, deleting duplicate records and samples containing missing values ​​in the transient operation data to obtain first intermediate data;

[0099] S23, calculating the mean μ2 and standard deviation σ2 of the first intermediate data, identifying data points that are beyond the range of the mean μ2 plus or minus three times the standard deviation σ2 according to the 3σ criterion, and removing these data points from the first intermediate data to obtain second intermediate data;

[0100] S24. Normalize the second intermediate data to obtain a test data set.

[0101] Further, in step S22, the mean μ2 and standard deviation σ2 are obtained by formulas (11) and (12), respectively, where a i is the data sample of the first intermediate data, n is the total sample size of the first intermediate data; the minimum value in the range of the mean plus or minus three times the standard deviation is min3σ2, and the maximum value is max3σ2, which are obtained by formulas (13) and (14) respectively.

[0102]

[0103] min3σ2=μ2-3σ2 Formula (13)

[0104] max3σ2=μ2+3σ2 Formula (14)

[0105] Preferably, step S24 specifically includes the following steps:

[0106] S240, respectively obtain the mean μ3 and standard deviation σ3 of the second intermediate data by formula (15) and (16), where b i is a data sample of the second intermediate data, and n is the total number of samples of the second intermediate data;

[0107]

[0108] S241. Normalize the second intermediate data using formula (17) to obtain a test data set, where B i is the normalized data.

[0109]

[0110] The effects of the present invention are described below in conjunction with specific embodiments; Specific embodiment 1

[0112] Based on the electromagnetic transient model of the direct-drive fan, a test data set of the electromagnetic transient model is obtained under three LVRT conditions, which is equivalent to the actual operation data of the direct-drive fan; then, based on the test data set and the identification method in the present invention, the LVRT control parameters of the direct-drive fan are obtained; then, the obtained LVRT control parameters of the direct-drive fan are substituted into the electromechanical transient model of the direct-drive fan, and the simulation of the same three conditions is completed to obtain the response curve, and finally the simulation results are compared with the test results in the test data set;

[0113] Three groups of LVRT working conditions are set, including high power, medium power and low power, and the parameters are shown in Table 1; then, based on the test data set and the identification method in the present invention, the LVRT control parameters of the direct-drive wind turbine are obtained; the LVRT control parameters are substituted into the electromechanical transient simulation model of the direct-drive wind turbine, and a three-phase symmetrical fault occurs at 4s and lasts for 0.625s. The voltage U and reactive current I of the grid-connected end of the direct-drive wind turbine under the three working conditions are obtained by simulation. q , the response curves of active power P and reactive power Q, and compare the simulation results of the three working conditions with the test results of the test data set, such as Figure 3 , Figure 4 and Figure 5 shown.

[0114] Table 1 Three working conditions

[0115]

[0116] from Figure 3 , Figure 4 and Figure 5 It can be seen that the simulation results obtained by substituting the LVRT control parameters obtained by the method of the present invention into the electromechanical transient model for simulation are highly fitted compared with the actual test results of the test data set, indicating that the identification method of the present invention can effectively identify the LVRT control parameters of the direct-drive wind turbine, verifying the accuracy and effectiveness of the identification method of the present invention. Specific embodiment 2

[0118] In order to illustrate the advantages of the improved vulture search algorithm used in the present invention, this specific embodiment will be compared with other algorithms, focusing on the comparison of convergence speed and convergence accuracy. This specific embodiment uses a particle swarm optimization algorithm (PSO) and a genetic algorithm (GA) to compare with the improved vulture search algorithm IBES in the present invention, and uses the same test data set and the same working conditions for testing. The direct-drive wind turbine LVRT control parameter identification results are obtained for each, as shown in Table 2. The root mean square error is used to measure the identification performance of the three algorithms. The number of iterations is selected as 50, and 50 simulations are performed. The average value is taken as the result of parameter identification, and the objective function convergence curves of the three different algorithms when solving the control parameters are obtained, as shown in Table 2. Figure 6 shown.

[0119] Table 2 Optimization results of three algorithms

[0120]

[0121] from Figure 6 It can be seen that the initial value of the PSO algorithm is larger, while the initial values ​​of the GA algorithm and the IBES algorithm are smaller. The size of the initial value reflects the distance of the algorithm from the optimal solution when it starts to iterate. The IBES algorithm has the smallest initial value, which means that it has better global search ability in the initial search stage and can approach the optimal solution faster. In terms of convergence speed, the IBES algorithm has the fastest convergence speed and can quickly converge to the optimal solution in a smaller number of iterations. The PSO algorithm converges faster in the early stage, but as the number of iterations increases, the convergence speed gradually slows down. The GA algorithm converges relatively slowly and requires more iterations to converge. In terms of convergence accuracy, the IBES algorithm has obvious advantages in optimization accuracy, and the final converged objective function value is the smallest, indicating that it can find better control parameters. The convergence accuracy of the PSO algorithm and the GA algorithm is relatively low, and the final converged objective function value is larger.

[0122] Therefore, the IBES algorithm has obvious advantages in solving the LVRT control parameters of direct-drive wind turbines. Compared with the PSO algorithm and the GA algorithm, it has a small initial value, fast convergence speed, and high convergence accuracy. It can effectively improve the identification accuracy and efficiency of the LVRT control parameters and is more suitable for solving complex optimization problems such as LVRT control parameters.

[0123] The above-mentioned method for identifying low voltage ride-through control parameters of direct-drive wind turbines based on the improved vulture search algorithm, on the one hand, the test data set used is initially collected by the direct-drive wind turbines under various LVRT conditions, and then repeated values, missing values, abnormal values ​​are eliminated, and normalized. Compared with the prior art, the adaptability to multiple conditions is enhanced, and the accuracy of LVRT control parameter identification is improved; on the other hand, in the identification process, the least squares method is used to obtain preliminary results of LVRT control parameters, which narrows the optimization range and improves the identification efficiency. The Bernoulli chaotic mapping is further used to improve the initial global search capability of the algorithm, and finally the adaptive weight ω is used to increase the optimization strength of the early stage of the identification algorithm to avoid falling into the local optimum, while reducing the optimization range of the late stage of the identification algorithm, thereby improving the accuracy of identification.

[0124] What is disclosed above is only a preferred embodiment of the present invention, which certainly cannot be used to limit the scope of rights of the present invention. A person skilled in the art can understand that all or part of the processes of the above embodiments and equivalent changes made according to the claims of the present invention still fall within the scope of the invention.

Claims

1. A method for identifying low voltage ride-through control parameters of a direct-drive wind turbine based on an improved vulture search algorithm, characterized in that: The following steps are involved: S1. Build the electromechanical transient model of the direct-drive wind turbine according to its physical characteristics, and determine the LVRT control parameters of the direct-drive wind turbine that need to be identified in the control mode under each control strategy; S2. Collect transient operation data of the direct-drive wind turbine under various LVRT conditions, and process the transient operation data as a test data set; S3. Based on the test data set obtained in S1, the improved vulture search algorithm IBES is used to identify the LVRT control parameters of the direct-drive wind turbine.

2. The method for identifying control parameters of low voltage ride-through of a direct-drive wind turbine based on an improved vulture search algorithm according to claim 1, characterized in that: Step S1 specifically includes the following steps: S10. According to the physical characteristics of the direct-drive wind turbine, an electromechanical transient model of the direct-drive wind turbine is constructed, the dynamic response process of the inner loop control and the phase-locked loop of the full power converter is ignored, and the operating characteristics of the LVRT process are considered; S11. According to the active power control strategy and reactive power control strategy of the direct-drive wind turbine converter, the specified current control mode under the two control strategies is analyzed: Among them, the specified current control mode under the active power control strategy is: Active current command I during LVRT pLVRT , obtained by formula (1), where K 1,Ip,LV and K 2,Ip,Lv All are low-power active current calculation coefficients; I pset It is the low active current setting value; I p0 is the initial active current; U t is the voltage drop degree; I pLVRT = K 1,Ip,LV U t + K 2,Ip,LV I p0 + I pset Equation (1) Specified current control mode under reactive power control strategy: Reactive current command I during LVRT qLVRT Obtained by formula (2), where K 1,Iq,LV and K 2,Iq,LV All are low-power reactive current calculation coefficients; U Lin is the low penetration threshold; I qset It is the setting value of low reactive current; I q0 is the initial reactive current; I qLVRT = K 1,Iq,LV (U Lin - U t ) + K 2,Iq,LV I q0 + I qset Equation (2) S12, according to the control logic of the specified current control mode under the two control strategies of the direct-drive wind turbine converter, Determine the LVRT control parameters of the direct-drive wind turbine that need to be identified, including: low-power active current calculation coefficient K 1,Ip,LV and K 2,Ip,Lv , low wear active current setting value I pset , low-power reactive current calculation coefficient K 1,Iq,LV and K 2,Iq,LV , low reactive current setting value I qset .

3. The method for identifying control parameters of low voltage ride-through of a direct-drive wind turbine based on an improved vulture search algorithm according to claim 1, characterized in that: Step S3 specifically includes the following steps: S30, based on the test data set, obtaining the initial value of the LVRT control parameter to be identified, and using the initial value of the LVRT control parameter as the initial data for IBES operation to determine the optimization range of the LVRT control parameter to be identified; S31. Construct the objective function J with the root mean square error, as shown in formula (3), where N is the number of test data sets, I p,e ,I q,e , P e , Q e are the deviations between the corresponding active current, reactive current, active power, reactive power data in the test data set and the identification data; S32, initializing basic iteration parameters, including population size, number of iterations and spatial dimension; S33. Use Bernoulli chaotic mapping to initialize the vulture population. The Bernoulli chaotic mapping is shown in formula (4), where μ1 is the chaos parameter with a value of 0.4, and y k is the current chaotic variable value, y k+1 is the updated chaotic variable value; S34, using formula (3) to calculate the fitness value of each vulture in the population, find the current optimal fitness value, that is, the minimum fitness value, and the corresponding position, set the search space with the position corresponding to the current optimal fitness value as the center, and update the position of each vulture according to the position corresponding to the current optimal fitness value, so that it moves toward the position corresponding to the optimal fitness value. The position of each vulture represents a set of LVRT control parameters. The position update formula is shown in formula (5), where P i , P new,i are the original position and updated position of the i-th vulture, i is a parameter that controls the position change, ranging from 1 to 2, r is a random value between 0 and 1, and P best is the position of the vulture with the best fitness value in the population, P mean is the position mean; P new,i =P best +sr(P mean -P i ) Formula (5) S35. From the individuals with better fitness values, select the position of the individuals as the search space. In the set search space, the vulture spirals in different directions. The update formulas of the spiral flight position are shown in formulas (6), (7) and (8), where P i , P new,i are the original position and updated position of the i-th vulture, respectively, P i+1 is the original position of the i+1th vulture; α is used to determine the search angle between the vulture and the center point, R rand It is used to determine the number of search cycles, and its value range is [0,1]; θ(i) and r(i) are the polar angle and polar diameter of each vulture's spiral flight, respectively, and y(i) and x(i) are the positions of the vulture in polar coordinates; P new,i = P i + y(i)(P i - P i+1 ) + x(i)(P i - P mean ) Equation (6) S36, the vulture swings from the current best position in the search space to the global best position, that is, for each vulture, according to the result of the fitness value, compare the advantages and disadvantages of its individual best position and the global best position; if the current position is better, use formulas (9) and (10) to update the individual best position; if there is a better position in the vulture population, then this better position is updated to the global best position; where c1 and c2 represent the vulture's movement towards the center point, that is, the position mean P mean and the global optimal position P mean The strength of the movement, the value range is [1,2]; ω is the adaptive weight; t is the number of iterations of the algorithm this time; P.S new,i ZR rand P.S best +ω[x(i)(P i -c1P mean )+y(i)(P i -c2P best )] house(9) S37, repeating steps S34, S35 and S36 until a predetermined number of iterations or an objective function is reached, that is, the fitness value is less than a given convergence error; S38, obtaining the final global optimal position, and extracting the direct-drive wind turbine LVRT control parameters that need to be identified therefrom.

4. The method for identifying control parameters of a direct-drive wind turbine low voltage ride-through based on an improved vulture search algorithm according to claim 3, characterized in that: In step S30, the least squares method is used to fit the initial value of the LVRT control parameter to be identified.

5. The method for identifying control parameters of low voltage ride-through of a direct-drive wind turbine based on an improved vulture search algorithm according to claim 1, characterized in that: Step S2 specifically includes the following steps: S20, set up multiple LVRT fault conditions for direct-drive fans, including high power, medium power and low power conditions; S21, collecting transient operation data of the direct-drive wind turbine under the above three LVRT conditions; S22, deleting duplicate records and samples containing missing values ​​in the transient operation data to obtain first intermediate data; S23, calculating the mean μ2 and standard deviation σ2 of the first intermediate data, identifying data points that are beyond the range of the mean μ2 plus or minus three times the standard deviation σ2 according to the 3σ criterion, and removing these data points from the first intermediate data to obtain second intermediate data; S24. Normalize the second intermediate data to obtain a test data set.

6. The method for identifying control parameters of low voltage ride through of a direct-drive wind turbine based on an improved vulture search algorithm according to claim 5, characterized in that: In step S22, the mean μ2 and standard deviation σ2 are obtained by formulas (11) and (12), respectively, where a i is the data sample of the first intermediate data, n is the total sample size of the first intermediate data; the minimum value in the range of the mean plus or minus three times the standard deviation is min3σ2, and the maximum value is max3σ2, which are obtained by formulas (13) and (14) respectively. min3σ2=μ2-3σ2 Formula (13) max3σ2=μ2+3σ2 Formula (14) 7. The method for identifying control parameters of low voltage ride-through of a direct-drive wind turbine based on an improved vulture search algorithm according to claim 5, characterized in that: Step S24 specifically includes the following steps: S240, respectively obtain the mean μ3 and standard deviation σ3 of the second intermediate data by formula (15) and (16), where b i is a data sample of the second intermediate data, and n is the total number of samples of the second intermediate data; S241. Normalize the second intermediate data using formula (17) to obtain a test data set, where B i is the normalized data.

Images