A method for optimizing a wind turbine yaw system start-up-to-wind control strategy

Abstract

The application discloses a kind of wind turbine yaw system start wind control strategy optimization method.The present application is extracted by operating data characteristics, according to the actual operating characteristics of wind turbine individual, the wind speed segmentation scheme of yaw system start wind control strategy is determined, the optimal start wind control parameter in each wind speed section is found using multi-objective optimization method, so as to effectively reduce the yaw number of wind turbine, while guaranteeing the power generation efficiency of wind turbine, improve the availability and economy of unit operation.

Description

[Technical Field]

[0001] This invention relates to the field of wind turbine generator optimization control technology, specifically to an optimization method for wind control strategy during the start-up of a wind turbine generator yaw system. [Background Technology]

[0002] With increasing public concern about environmental issues and a growing demand for green energy worldwide, wind energy, as a green energy source with enormous resource potential and relatively mature applications, is becoming increasingly important. Wind turbines are key equipment for wind energy utilization, and improving their power generation efficiency and reducing failure rates are focal points for the wind power industry. The yaw system is a crucial subsystem of a wind turbine, its main function being to adjust the turbine's angle of attack to effectively track wind direction. The performance of the yaw system is a key factor affecting the power generation efficiency of the wind turbine. Clearly, a precise and rapid response to any change in wind direction is most beneficial for improving instantaneous power generation efficiency. However, frequent yaw system starts can lead to high loads, causing failures or malfunctions in critical components such as the yaw gear and slewing bearing, adversely affecting the lifespan and safety of the wind turbine. There is a mutually restrictive relationship between the number of yaw cycles and power generation efficiency, making it difficult to simultaneously guarantee efficiency in areas with complex wind conditions.

[0003] Most large wind turbines employ automatic yaw control. Currently, the practical yaw start-up wind control strategy is relatively simple: within the entire operating wind speed range, when the yaw error angle exceeds its allowable range and the cumulative duration exceeds the delay time threshold, the wind turbine initiates yaw control. Research on this strategy, both domestically and internationally, is relatively limited. Three approaches have been proposed: 1) Dividing the wind speed range above the cut-off wind speed into multiple wind speed intervals based on the wind speed probability distribution curve, and employing different yaw control parameters within each interval; 2) Optimizing the control parameters using a bacterial community drug-attraction optimization algorithm to enhance their adaptability; 3) Proposing a wind speed segmented yaw control strategy based on the force analysis results of the yaw state, with different empirically chosen values ​​for the yaw control parameters in different wind speed segments. Simulation results show that this strategy can slightly reduce the number of yaw system actions without affecting power generation. Therefore, dividing the wind speed range where the wind turbine operates into several sub-intervals and employing different control parameters for each is one of the development directions for yaw system start-up wind control. However, the wind speed range division in the above methods is achieved through subjective analysis and does not form an objective standard for wind speed range division, which is not conducive to its widespread application. The wind speed range division only considers the mechanical characteristics of the wind turbine or the statistical characteristics of wind resources in the wind farm, without fully analyzing the actual operating characteristics of individual wind turbines. Furthermore, these methods transform the multi-objective optimization problem of yaw control optimization into a single-objective optimization problem, failing to effectively balance the two objectives of reducing yaw frequency and ensuring the power generation efficiency of the wind turbine based on specific circumstances.

[0004] It should also be noted that when operating in the wind speed range above the rated wind speed, large wind turbines do not need to fully utilize wind energy. In this case, using a larger yaw error angle threshold and a longer delay time threshold is sufficient to suppress frequent yaw system actions. Therefore, the wind control strategy for starting the yaw system of large wind turbines that needs optimization is mainly the wind control strategy used when entering the wind speed range above the rated wind speed. [Summary of the Invention]

[0005] To address the aforementioned technical problems, this invention provides an optimization method for wind control strategy during the startup of a wind turbine yaw system.

[0006] This invention is achieved through the following technical solution:

[0007] An optimization method for wind turbine yaw system startup wind control strategy, comprising the following steps:

[0008] S1: Read the historical operation data of the target wind turbine after it has been stably in service for hy years with the first preset time as the time resolution from the SCADA system to construct dataset Cm; and the historical operation data of the most recent consecutive hs months with the second preset time as the time resolution to construct dataset Cs;

[0009] S2: Cut the target wind turbine unit into wind speed v in and rated wind speed v rated The wind speed range between [v] in v rated ]Discrete into ZR subintervals at equal intervals [v in , v1), [v1, v2),…, [v ZR-1 v ZR ];

[0010] S3: Using the data in Cm, based on the extracted wind speed range yaw comprehensive characteristics, determine [v in v rated Wind speed segmentation scheme for activating the yaw system's counter-wind strategy within a wind speed range;

[0011] S4: If there are NI wind speed intervals for which the yaw system activation strategy needs to be independently set, as obtained in step S3, with the objectives of minimizing the number of yaws Q and maximizing the power generation W, a multi-objective optimization algorithm is used to set the yaw error angle threshold θ for each wind speed interval using the dataset Cs. i Delay time threshold T i Optimize (i = 1, 2, ..., NI) to obtain the Pareto front solution set of the yaw control parameters;

[0012] S5: Determine the importance weights of the two optimization objectives, namely, the number of yaws and the amount of power generated, and use the TOPSIS method of multi-attribute decision analysis to obtain the optimal solution of the yaw start control parameters from the Pareto front solution set;

[0013] S6: Set the yaw start control strategy for each interval as follows: when the yaw error angle in the i-th (i = 1, 2, ..., NI) wind speed interval that requires independent setting of the yaw system start wind strategy is greater than the threshold θ i And the duration exceeds the delay time threshold T i At that time, start the wind countermeasures.

[0014] The method for optimizing wind control strategy during the start-up of a wind turbine yaw system, as described above, includes the following steps in step S3: the comprehensive yaw characteristics are calculated and synthesized from the "yaw error-power" correlation characteristics and wind speed turbulence characteristics. The method for extracting the "yaw error-power" correlation characteristics for wind speed sub-intervals includes the following steps:

[0015] A1: Calculate the baseline power generation for wind speed range j. PS:

[0016]

[0017] Where sp represents the sample points in dataset Cm; SCm is a subset of sample points in dataset Cm whose wind speed is within the current wind speed sub-interval and whose absolute value of the yaw error angle parameter is within the range of [0°, 1°); sp power The active power parameter of the sample point sp is the output power parameter; NSCm is the number of sample points in the subset SCm.

[0018] A2: Calculate the power generation PB of the deviation in wind speed range j:

[0019]

[0020] Where BCm is a subset of sample points in dataset Cm whose wind speed parameters are within the current wind speed sub-interval and whose absolute values ​​of yaw error angle parameters are within the range [Fb°, Fb+1°), where 4≤Fb h ≤10; sp power is the output active power parameter of sample point sp; NBCm is the number of sample points in subset BCm;

[0021] A3: Calculate the "yaw error - power" correlation characteristic F for this wind speed range. ya-p :

[0022]

[0023] Wherein, PD is the power reference constant, and PD takes values ​​in the interval [100, 1000].

[0024] The optimization method for wind control strategy during the start-up of a wind turbine yaw system, as described above, involves calculating the comprehensive yaw characteristics based on the correlation characteristics of "yaw error-power" and wind speed turbulence characteristics in each wind speed sub-interval.

[0025] F c =bp×F ya-p +(1-bp)×F ti

[0026] Among them, F c That is, the yaw comprehensive feature, where bp is the weight of the "yaw error-power" related feature, and bp takes a value in the interval [0.7, 0.95].

[0027] The method for optimizing wind control strategy during the start-up of a wind turbine yaw system, as described above, involves determining [v] in step S3 based on the comprehensive yaw characteristics of each wind speed sub-interval. in v ratedThe method for initiating a yaw system's counter-wind strategy within a wind speed range, according to the wind speed segmentation scheme, includes the following steps:

[0028] B1: Let j = 1, calculate the yaw comprehensive characteristics of the j-th wind speed sub-interval.

[0029] B2: Let i = 1, take the j-th wind speed sub-interval as the i-th wind speed segment that requires independent yaw system activation and wind countermeasure strategy, and take the comprehensive yaw characteristics of this wind speed interval as the basis. The comprehensive yaw characteristics of the i-th wind speed segment where the yaw system needs to be independently configured to initiate the wind countermeasure strategy.

[0030] B3: Let j = j + 1, calculate the yaw characteristics of the j-th wind speed sub-interval.

[0031] B4: If If true, the j-th wind speed sub-interval will be merged into the i-th wind speed segment that requires independent yaw system activation and wind countermeasure strategy. The intervals will then be recalculated.

[0032] Otherwise, let i = i + 1, and take the j-th wind speed sub-interval as the i-th wind speed segment that requires independent setting of the yaw system to initiate the wind countermeasure strategy, and at the same time... As a new

[0033] B5: Determine if the j-th wind speed subinterval is the last discrete wind speed subinterval; if so, the wind speed segmentation ends, and [v] is obtained. in v rated If the wind speed range is within the specified range, set up the yaw system to initiate the wind countermeasure strategy in a segmented wind speed scheme; otherwise, proceed to step B3.

[0034] In the wind control strategy optimization method for the start-up of a wind turbine yaw system as described above, in step S5:

[0035] The formula for calculating the objective Q during the optimization process is: Q = ∑Sa, where Sa is the number of times the yaw start strategy is satisfied in the analyzed running data;

[0036] The calculation formula for the target W during the optimization process is: W=∑P×Δt, where P is the instantaneous power of the wind turbine, Δt is the data recording time interval of data Cs, and the value range of Δt is [0.5, 5].

[0037] The formula for calculating the instantaneous power P of a wind turbine is: P = Co × ω 3 ×cos 3θ+Cp, where θ is the yaw error angle of the wind turbine, ω is the wind turbine speed, Co is a coefficient related to the wind turbine diameter, air density, and wind turbine utilization coefficient, and Cp is a compensation coefficient related to the error and generator conversion efficiency.

[0038] The above-described optimization method for wind control strategy during the start-up of a wind turbine yaw system requires the following steps to calculate the parameters Co and Cp of the instantaneous power P. Before using the power calculation formula, these parameters need to be solved using the least squares method with the yaw error angle, rotor speed, and output active power parameters recorded in the dataset Cs.

[0039] C1: Extract the wind speed in Cs at [v] in v d Sample points within the range 6≤v d For wind speeds ≤9, the least squares method is used to solve for Co and Cp within this wind speed range using the above sample points, and they are denoted as Co1 and Cp1, respectively.

[0040] C2: Extract the wind speed in Cs at (v) d v rated Within the range of sample points, the least squares method is used to solve for Co and Cp within this wind speed range, which are denoted as Co2 and Cp2, respectively.

[0041] C3:

[0042]

[0043]

[0044] Co and Cp are calculated using the formulas above.

[0045] The wind control strategy optimization method for the yaw system startup of a wind turbine, as described above, employs a second-generation non-dominated sorting genetic algorithm (NSGA-II) in step S5. This algorithm is used to determine the yaw error angle threshold θ for each wind speed range. k and delay time threshold T k The method for optimizing to (K = 1, 2, ..., IN) includes the following steps:

[0046] D1: Set the yaw error angle threshold θ to be optimized. k The reasonable range of values ​​is limited to [0°, 60°], and the delay time threshold T k The reasonable range is limited to [0s, 360s];

[0047] D2: Following the NSGA-II algorithm, the optimal individual coefficient is set to Ns. Using Popps as the population size, create variables to be optimized [θ1, θ2, ..., θ]. inT1, T2, ..., T IN The initial population; where 0.5≤Ns≤0.9, 50≤Pops≤300;

[0048] D3: Optimize the calculation methods of objective Q and W according to the NSGA-II algorithm and the yaw start wind strategy, evaluate the fitness of each individual in the population, perform non-dominated ranking of the population according to the dominance relationship between individuals, and divide each individual into different levels;

[0049] D4: Calculate the crowding distance of individuals in each level according to the NSGA-II algorithm, and select the next generation of individuals based on the non-dominated ranking and crowding distance;

[0050] D5: The crossover rate is CR and the mutation rate is MR. Crossover and mutation operations are applied to the selected individuals to generate the next generation population; where 0.5≤CR≤0.95 and 0.05≤MR≤0.55;

[0051] D6: If the maximum number of generations MI is reached, output the Pareto front solution set of the current population as the yaw control parameter; otherwise, go to step D3; where 30≤MI≤100.

[0052] In the wind turbine yaw system startup optimization method described above, in step S5, the yaw error angle θ update formula during multi-objective optimization is as follows:

[0053]

[0054] Where D is the wind direction value at the current moment; D -1 It is the wind direction value at the previous moment; θ -1 VY is the yaw error angle at the previous moment, and VY is the yaw speed, which is 0.3deg / s≤VY≤1.2deg / s.

[0055] The method for optimizing wind control strategy during the start-up of a wind turbine yaw system, as described above, includes the following steps in step S5: The method for determining the optimal solution from the Pareto solution set using the TOPSIS method:

[0056] E1: Positively process the yaw rate index for each scheme: PQ k =max(Q k )-Q k k = 1, 2, ..., NP

[0057] Among them, Q k PQ is the yaw rate index for the k-th solution in the Pareto solution set. kNP is the number of yaws after positive transformation, and NP is the number of solutions in the Pareto front solution set. According to the principle of the NSGA-II algorithm, NP = Ns × Pops.

[0058] E2: Construct the standardized evaluation index vector Z for each solution in the solution set. k Where k = 1, 2, ..., NP:

[0059]

[0060] E3: Calculate the evaluation index vector Z of the ideal solution. + The evaluation index vector Z of the negative ideal solution - :

[0061]

[0062]

[0063] E4: Calculate the distance D from each solution to the ideal solution and the negative ideal solution. k + D k - :

[0064]

[0065]

[0066] Wherein, wt1 and wt2 are the importance weights of the power generation index and the number of yaws, respectively, and the value range of wt1 is [0.6, 0.8] and wt1+wt2=1;

[0067] E5: Calculate the score Sr for each solution. k :

[0068] E6: Rank the solutions according to their scores and select the solution with the highest score as the optimal solution.

[0069] As described above, in the method for optimizing wind control strategy during the start-up of a wind turbine yaw system, in step S2, the target wind turbine's cut-in wind speed v in The value range is [2m / s, 6m / s]; rated wind speed v rated The value range is [9m / s, 16m / s]; the number of discrete sub-wind speed intervals ZR is a positive integer, with a value range of [6, 20].

[0070] Compared with the prior art, this application has the following advantages:

[0071] This invention discloses an optimization method for wind turbine yaw system startup wind control strategy. Through operational data feature extraction, it determines the wind speed segmentation scheme for the yaw system startup wind control strategy based on the actual operating characteristics of each individual wind turbine. A multi-objective optimization method is employed to find the optimal startup wind control parameters within each wind speed segment. This effectively reduces the number of yaw cycles while ensuring the wind turbine's power generation efficiency, improving the availability and economy of the unit's operation. An objective wind speed range division standard is proposed, which is easy to implement and facilitates the widespread application of the method. Wind speed segmentation is based on the proposed comprehensive yaw feature, which considers both the impact of the yaw error angle measured during actual wind turbine operation on the wind turbine's power generation efficiency and the stability of wind resources; the basis for wind speed segmentation is scientific and comprehensive. When searching for the optimal start-up wind control parameters within each wind speed range using multi-objective optimization, the accuracy of calculating one of the optimization objectives, power generation W, depends on the accuracy of solving for the two constant coefficients in the formula for calculating the instantaneous power P of the wind turbine. Based on extensive analysis, this invention proposes a constant coefficient solution method that identifies parameters in two intervals according to wind speed, improving the accuracy of power generation optimization calculation. Based on the NSGA-II algorithm and the TOPSIS method, a complete multi-objective decision-making process for the start-up wind control of the wind turbine yaw system is presented. Compared with optimization methods that transform multi-objective problems into single-objective problems, this method can better balance the two objectives of effectively reducing yaw frequency and ensuring wind turbine power generation efficiency based on specific circumstances. This invention can also be extended to multi-objective decision optimization of other complex electromechanical control systems. [Attached Image Description]

[0072] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below.

[0073] Figure 1 This is a flowchart illustrating the overall optimization process of the present invention;

[0074] Figure 2 In this embodiment of the invention, the wind turbine's cut-in wind speed to the rated fraction is discretized at intervals of 1 m / s, and the least squares method is used to solve for Co in each discrete wind speed interval to obtain the result.

[0075] Figure 3 This is a flowchart illustrating the wind speed segmentation scheme for determining the yaw system activation wind-fighting strategy in an embodiment of the present invention.

[0076] Figure 4 This is a flowchart illustrating the multi-objective optimization of yaw start control parameters according to an embodiment of the present invention.

[0077] Figure 5 This is a Pareto front diagram for multi-objective optimization in an embodiment of the present invention;

[0078] Figure 6 Here is the yaw error angle variation curve from a short-timescale simulation of an embodiment of the present invention:

[0079] (a) The first day of the sampling;

[0080] (b) The day after the sampling;

[0081] (c) The third day after sampling;

[0082] (d) The fourth day after sampling;

[0083] Figure 7 Here is the yaw rate variation curve from a short-timescale simulation of an embodiment of the present invention:

[0084] (a) The first day of the sampling;

[0085] (b) The day after the sampling;

[0086] (c) The third day after sampling;

[0087] (d) The fourth day of sampling.

Detailed Implementation Methods

[0088] To make the technical problems solved, the technical solutions, and the beneficial effects of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0089] An optimization method for wind turbine yaw system startup wind control strategy, comprising the following steps:

[0090] S1: Read the historical operation data of the target wind turbine after it has been stably in service for hy years with the first preset time as the time resolution from the SCADA system to construct dataset Cm; and the historical operation data of the most recent consecutive hs months with the second preset time as the time resolution to construct dataset Cs;

[0091] S2: Cut the target wind turbine unit into wind speed v in and rated wind speed v rated The wind speed range between [v] in v rated ]Discrete into ZR subintervals at equal intervals [v in , v1), [v1, v2),…, [v ZR-1 v ZR ];

[0092] S3: Using the data in Cm, based on the extracted wind speed range yaw comprehensive characteristics, determine [vin v rated Wind speed segmentation scheme for activating the yaw system's counter-wind strategy within a wind speed range;

[0093] S4: If there are NI wind speed intervals for which the yaw system activation strategy needs to be independently set, as obtained in step S3, with the objectives of minimizing the number of yaws Q and maximizing the power generation W, a multi-objective optimization algorithm is used to set the yaw error angle threshold θ for each wind speed interval using the dataset Cs. i Delay time threshold T i Optimize (i = 1, 2, ..., NI) to obtain the Pareto front solution set of the above yaw control parameters;

[0094] S5: Determine the importance weights of the two optimization objectives, namely, the number of yaws and the amount of power generated, and use the TOPSIS method of multi-attribute decision analysis to obtain the optimal solution of the yaw start control parameters from the Pareto front solution set;

[0095] S6: Set the yaw start control strategy for each interval as follows: when the yaw error angle in the i-th (i = 1, 2, ..., NI) wind speed interval that requires independent setting of the yaw system start wind strategy is greater than the threshold θ i And the duration exceeds the delay time threshold T i At that time, start the wind countermeasures.

[0096] Furthermore, as a preferred embodiment of this solution and not a limitation, the first preset time is 10 minutes, and the second preset time is 1 second. The SCADA system is a Supervisory Control and Data Acquisition system.

[0097] Furthermore, as a preferred embodiment of this solution and not a limitation, in step S1, the parameters included in the sample points in the datasets Cm and Cs include the current time, wind speed, wind direction, yaw error angle, and output power, where 1≤hy≤3 and 3≤hs≤12.

[0098] Furthermore, as a preferred embodiment of this solution and not a limitation, in step S3, the yaw composite feature is calculated and synthesized from the "yaw error-power" correlation feature and the wind speed turbulence feature. The method for extracting the "yaw error-power" correlation feature for wind speed sub-intervals includes the following steps:

[0099] A1: Calculate the baseline power generation for wind speed range j. PS:

[0100]

[0101] Where sp represents the sample points in dataset Cm; SCm is a subset of sample points in dataset Cm whose wind speed is within the current wind speed sub-interval and whose absolute value of the yaw error angle parameter is within the range of [0°, 1°); sp power The output active power parameter of sample point sp; NSCm is the number of sample points in subset SCm:

[0102] A2: Calculate the power generation PB of the deviation in wind speed range j:

[0103]

[0104] Where BCm is a subset of sample points in dataset Cm whose wind speed parameters are within the current wind speed sub-interval and whose absolute values ​​of yaw error angle parameters are within the range [Fb°, Fb+1°), where 4≤Fb h ≤10; sp power is the output active power parameter of sample point sp; NBCm is the number of sample points in subset BCm;

[0105] A3: Calculate the "yaw error - power" correlation characteristic F for this wind speed range. ya-p :

[0106]

[0107] Wherein, PD is the power reference constant, and PD takes values ​​in the interval [100, 1000].

[0108] Furthermore, as a preferred embodiment of this solution and not a limitation thereof, the [v] in v rated In step S4 of the internal wind speed segmentation, Δ is the interval difference threshold, with a value range of [0.1, 0.4].

[0109] Furthermore, as a preferred embodiment of this solution and not a limitation, the wind speed turbulence characteristics F of the wind speed sub-intervals are... ti The standard deviation of turbulence intensity was calculated using data from dataset Cm according to the standard deviation method for turbulence intensity calculation as specified in IEC 61400-1-2019 (International Electrotechnical Standard).

[0110] Furthermore, as a preferred, but not limited, implementation of this scheme, the method for calculating the comprehensive yaw characteristic from the "yaw error-power" correlation characteristics and wind speed turbulence characteristics of each wind speed sub-interval is as follows:

[0111] F c =bp×F ya-p +(1-bp)×F ti

[0112] Among them, F cThat is, the yaw comprehensive feature, where bp is the weight of the "yaw error-power" related feature, and bp takes a value in the interval [0.7, 0.95].

[0113] Furthermore, as a preferred embodiment of this solution and not a limitation, in step S3, based on the comprehensive yaw characteristics of each wind speed sub-interval, [v in v rated The method for initiating a yaw system's counter-wind strategy within a wind speed range, according to the wind speed segmentation scheme, includes the following steps:

[0114] B1: Let j = 1, calculate the yaw comprehensive characteristics of the j-th wind speed sub-interval.

[0115] B2: Let i = 1, take the j-th wind speed sub-interval as the i-th wind speed segment that requires independent yaw system activation and wind countermeasure strategy, and take the comprehensive yaw characteristics of this wind speed interval as the basis. The comprehensive yaw characteristics of the i-th wind speed segment where the yaw system needs to be independently configured to initiate the wind countermeasure strategy.

[0116] B3: Let j = j + 1, calculate the yaw characteristics of the j-th wind speed sub-interval.

[0117] B4: If If true, the j-th wind speed sub-interval will be merged into the i-th wind speed segment that requires independent yaw system activation and wind countermeasure strategy. The intervals will then be recalculated.

[0118] Otherwise, let i = i + 1, and take the j-th wind speed sub-interval as the i-th wind speed segment that requires independent setting of the yaw system to initiate the wind countermeasure strategy, and at the same time... As a new

[0119] B5: Determine if the j-th wind speed subinterval is the last discrete wind speed subinterval; if so, the wind speed segmentation ends, and [v] is obtained. in v rated If the wind speed range is within the specified range, set up the yaw system to initiate the wind countermeasure strategy in a segmented wind speed scheme; otherwise, proceed to step B3.

[0120] Furthermore, as a preferred embodiment of this solution and not a limitation, in step S5:

[0121] The formula for calculating the objective Q during the optimization process is: Q = ∑Sa, where Sa is the number of times the yaw start strategy is satisfied in the analyzed running data;

[0122] Furthermore, as a preferred implementation of this solution and not a limitation, the formula for calculating the target W during the optimization process is: W = ∑P × Δt, where P is the instantaneous power of the wind turbine, Δt is the data recording time interval of data Cs, and the value range of Δt is [0.5, 5].

[0123] The formula for calculating the instantaneous power P of a wind turbine is: P = Co × ω 3 ×cos 3 θ+Cp, where θ is the yaw error angle of the wind turbine, ω is the wind turbine speed, Co is a coefficient related to the wind turbine diameter, air density, and wind turbine utilization coefficient, and Cp is a compensation coefficient related to the error and generator conversion efficiency.

[0124] Furthermore, as a preferred, but not limited, implementation of this scheme, the parameters Co and Cp for calculating the instantaneous power P need to be solved using the least squares method before using the power calculation formula, based on the yaw error angle, rotor speed, and output active power parameters recorded in the dataset Cs. This includes the following steps:

[0125] C1: Extract the wind speed in Cs at [v] in v d Sample points within the range 6≤v d For wind speeds ≤9, the least squares method is used to solve for Co and Cp within this wind speed range using the above sample points, and they are denoted as Co1 and Cp1, respectively.

[0126] C2: Extract the wind speed in Cs at (v) d v rated Within the range of sample points, the least squares method is used to solve for Co and Cp within this wind speed range, which are denoted as Co2 and Cp2, respectively.

[0127] C3:

[0128]

[0129]

[0130] Co and Cp are calculated using the formulas above.

[0131] Furthermore, as a preferred embodiment of this solution and not a limitation, in step S5, the multi-objective optimization algorithm used is the second-generation non-dominated sorting genetic algorithm, namely the NSGA-II algorithm, which sets the yaw error angle threshold θ for each wind speed range. k and delay time threshold T k The method for optimizing to (K = 1, 2, ..., IN) includes the following steps:

[0132] D1: Set the yaw error angle threshold θ to be optimized. kThe reasonable range of values ​​is limited to [0°, 60°], and the delay time threshold T k The reasonable range is limited to [0s, 360s];

[0133] D2: Following the NSGA-II algorithm, the optimal individual coefficient is set to Ns. Using Popps as the population size, create variables to be optimized [θ1, θ2, ..., θ]. in T1, T2, ..., T IN The initial population; where 0.5≤Ns≤0.9, 50≤Pops≤300;

[0134] D3: Optimize the calculation methods of objective Q and W according to the NSGA-II algorithm and the yaw start wind strategy, evaluate the fitness of each individual in the population, perform non-dominated ranking of the population according to the dominance relationship between individuals, and divide each individual into different levels;

[0135] D4: Calculate the crowding distance of individuals in each level according to the NSGA-II algorithm, and select the next generation of individuals based on the non-dominated ranking and crowding distance;

[0136] D5: The crossover rate is CR and the mutation rate is MR. Crossover and mutation operations are applied to the selected individuals to generate the next generation population; where 0.5≤CR≤0.95 and 0.05≤MR≤0.55;

[0137] D6: If the maximum number of generations MI is reached, output the Pareto front solution set of the current population as the yaw control parameter; otherwise, go to step D3; where 30≤MI≤100.

[0138] Furthermore, as a preferred embodiment of this solution and not a limitation, in step S5, the formula for updating the flight error angle θ when performing multi-objective optimization is:

[0139]

[0140] Where D is the wind direction value at the current moment; D -1 It is the wind direction value at the previous moment; θ -1 VY is the yaw error angle at the previous moment, and VY is the yaw speed, which is 0.3deg / s≤VY≤1.2deg / s.

[0141] Furthermore, as a preferred embodiment of this solution and not a limitation thereof, step S5, the method for determining the optimal solution from the Pareto solution set using the TOPSIS method, includes the following steps:

[0142] E1: Positively process the yaw rate index for each scheme: PQ k =max(Q k)-Q k k = 1, 2, ..., NP

[0143] Among them, Q k PQ is the yaw rate index for the k-th solution in the Pareto solution set. k NP is the number of yaws after positive transformation, and NP is the number of solutions in the Pareto front solution set. According to the principle of the NSGA-II algorithm, NP = Ns × Pops.

[0144] E2: Construct the standardized evaluation index vector Z for each solution in the solution set. k Where k = 1, 2, ..., NP:

[0145]

[0146] E3: Calculate the evaluation index vector Z of the ideal solution. + The evaluation index vector Z of the negative ideal solution - :

[0147]

[0148]

[0149] E4: Calculate the distance D from each solution to the ideal solution and the negative ideal solution. k + D k - :

[0150]

[0151]

[0152] Wherein, wt1 and wt2 are the importance weights of the power generation index and the number of yaws, respectively, and the value range of wt1 is [0.6, 0.8] and wt1+wt2=1;

[0153] E5: Calculate the score Sr for each solution. k :

[0154] E6: Rank the solutions according to their scores and select the solution with the highest score as the optimal solution.

[0155] Furthermore, as a preferred embodiment of this solution and not a limitation, in step S2, the target wind turbine's cut-in wind speed v in The value range is [2m / s, 6m / s].

[0156] Furthermore, as a preferred embodiment of this solution and not a limitation, the rated wind speed v ratedThe value range is [9m / s, 16m / s]; the number of discrete sub-wind speed intervals ZR is a positive integer, with a value range of [6, 20].

[0157] In this embodiment, the data comes from a large direct-drive permanent magnet wind turbine operating in a mountainous wind farm in southern China, with a cut-in wind speed v. in The value is 3.0 m / s, and the rated wind speed is v. rated The value is 11 m / s.

[0158] An optimization method for wind turbine yaw system startup wind control strategy, comprising the following steps:

[0159] S1: Read the historical operating data recorded in minutes for 1 year (hy=1) after the target wind turbine has been stably in service from the SCADA system to construct dataset Cm; and the historical operating data recorded in seconds for the most recent 4 consecutive months (hs=4) to construct dataset Cs; the parameters included in the sample points in datasets Cm and Cs include the current time, wind speed, wind direction, yaw error angle, and output power.

[0160] S2: Cut the target wind turbine unit into wind speed v in and rated wind speed v rated The wind speed range [3, 11] is discretized at equal intervals into 8 (ZR = 8) sub-intervals [v in , v1), [v1, v2),…, [v ZR-1 v ZR ]; Determine [v in v rated The method for initiating a yaw system's counter-wind strategy within a wind speed range, according to the wind speed segmentation scheme, includes the following steps:

[0161] B1: Let j = 1, calculate the yaw comprehensive characteristics of the j-th wind speed sub-interval. During this process, the correlation characteristic F of "yaw error-power" is calculated. ya-p At that time, Fb h The value is 8, and the power reference constant PD is 200: wind speed turbulence characteristic F tiThe standard deviation method for turbulence intensity calculation is performed using the dataset Cs according to the IEC 61400-1-2019 standard. For details, please refer to the reference [IEC 61400-1.2019. Wind energy generation system - Part I. Design requirements: Geneva, Switzerland.]. When calculating the comprehensive yaw characteristics from the "yaw error-power" correlation characteristics and wind speed turbulence characteristics of each wind speed sub-interval, the weight bp of the "yaw error-power" correlation characteristics is set to 0.8.

[0162] B2: Let i = 1, take the j-th wind speed sub-interval as the i-th wind speed segment that requires independent yaw system activation and wind countermeasure strategy, and take the comprehensive yaw characteristics of this wind speed interval as the basis. The comprehensive yaw characteristics of the i-th wind speed segment where the yaw system needs to be independently configured to initiate the wind countermeasure strategy.

[0163] B3: Let j = j + 1, calculate the yaw characteristics of the j-th wind speed sub-interval. In this process, the parameter values ​​are the same as in step B1.

[0164] B4: If If true, the j-th wind speed sub-interval will be merged into the i-th wind speed segment that requires independent yaw system activation and wind countermeasure strategy. The intervals will then be recalculated.

[0165] Otherwise, let i = i + 1, and take the j-th wind speed sub-interval as the i-th wind speed segment that requires independent setting of the yaw system to initiate the wind countermeasure strategy, and at the same time... As a new In this step, the interval difference threshold Δ is set to 0.14. The parameter values ​​for the yaw composite feature are the same as in step B1.

[0166] B5: Determine if the j-th wind speed subinterval is the last discrete wind speed subinterval; if so, the wind speed segmentation ends, and [v in v rated If the wind speed range is within the specified range, set the yaw system to initiate the wind countermeasure strategy in a segmented wind speed configuration; otherwise, proceed to step B3.

[0167] In this embodiment, following the method described above, [v in v ratedThe system should be divided into three intervals (NI=3) for setting the yaw system to initiate wind countermeasures: [3m / s, 6m / s), [6m / s, 8m / s), and [8m / s, 11m / s]. The final calculated yaw composite characteristic for the first wind speed sub-interval is 0.12; for the second wind speed sub-interval, it is 0.45; and for the third wind speed sub-interval, it is 0.08.

[0168] After obtaining the segmented wind speed scheme for setting up the yaw system and initiating the wind countermeasure strategy within the wind speed range, before optimizing the yaw control parameters for each wind speed range using a multi-objective optimization algorithm, it is necessary to solve for the two constant coefficients Co and Cp in the formula for calculating the instantaneous power P of the wind turbine (instantaneous power P is the basis for calculating one of the optimization objectives, power generation W). Among these two constant coefficients, the accuracy of the proportional coefficient Co has a much greater impact on the accuracy of the instantaneous power P calculation result than the constant term coefficient Cp. Figure 2 The wind turbine's cut-in wind speed to its rated fraction is discretized at intervals of 1 m / s, and the least squares method is used to solve for Co within each discrete wind speed interval. As shown in the figure, the values ​​of Co are relatively close when the wind speed is less than 7 m / s; the values ​​of Co are also relatively close when the wind speed is greater than 7 m / s; however, the values ​​of Co differ significantly between the two intervals. Therefore, the method proposed in this invention solves for Co and Cp in two segments, ensuring the accuracy of power P calculation during the optimization process without making the calculation process overly complex. Figure 3 Based on the analysis results, in this embodiment, the critical wind speed v with constant coefficients is solved piecewise. d The value is 7 m / s.

[0169] Next, with the objectives of minimizing the number of yaws Q and maximizing the power generation W, the NSGA-II multi-objective optimization algorithm was used with dataset Cs to optimize the parameter group [θ1, θ2, θ3, T1, T2, T3] consisting of the yaw error angle threshold and delay time threshold parameters for each wind speed range. This process is detailed in the appendix. Figure 4 The optimal individual coefficient Ns, population size Popps, crossover rate CR, mutation rate MR, and maximum number of generations MI were set to 0.7, 50, 0.8, 0.2, and 30, respectively. Referring to the wind control strategy implemented by this wind turbine, the yaw speed VY was set to 0.6 deg / s. The Pareto front plot of the optimal solution set for this multi-objective optimization is shown in the appendix. Figure 5 .

[0170] After obtaining the optimal solution set, the power generation weight wt1 was set to 0.7, and the yaw count wt2 was set to 0.3. The comprehensive optimal solution obtained by the TOPSIS method is shown in Table 1 (that is, the optimized start-up yaw control parameters of the wind turbine for the wind speed range above the cut-in wind speed and below the rated wind speed). Table 1 also includes the yaw control parameters originally used by the wind turbine (the cut-out wind speed of the wind turbine is 25 m / s, and the same set of start-up yaw control parameters are used for the entire operating wind speed range from the cut-in wind speed to the cut-out wind speed).

[0171] Table 1. Yaw start control parameters originally used for this wind turbine and yaw start control parameters obtained by this invention.

[0172]

[0173] To verify the effectiveness of this invention, a comparative simulation was conducted between the optimized yaw system startup wind control strategy of this invention and the original startup wind control strategy used by the wind turbine. The simulation method was as follows: wind direction and speed data recorded at a time resolution of 1 second, read from the wind turbine's SCADA system, were used as simulation input; whether the wind turbine started yaw was determined according to the two yaw system startup wind control strategies (this invention does not optimize the yaw startup control parameters above the rated wind speed; therefore, when the wind speed is greater than the wind turbine's rated wind speed, the optimized yaw startup wind control strategy uses the original yaw startup control parameters of the wind turbine); the yaw error angle update method of the wind turbine was the same as the update method in the optimization process; the calculation method of the wind turbine's power generation was the same as the calculation method of the optimized target power generation W in the optimization process.

[0174] First, wind direction and speed data recorded for four days at a time resolution of 1 second were randomly selected from the wind turbine's SCADA system (the time interval between the daily data was more than 10 days). The two yaw start-up split control strategies were compared from a short-term time scale perspective. During the simulation, the yaw error angle variation curve is shown below. Figure 6 As shown, the variation in the number of yaws is as follows: Figure 7 As shown in Table 2, the power generation is as follows.

[0175] Table 2 Comparison of power generation in short-timescale simulations

[0176]

[0177] Depend on Figure 6 , Figure 7As shown in Table 2, the optimized yaw start-up wind control strategy significantly reduces the number of yaws. The yaw error angle of wind turbines with wind speeds in the range of 3.0 m / s to 6.0 m / s increases overall, while the yaw error angle with wind speeds above 6 m / s shows no significant change compared to the original yaw start-up wind control strategy. The increased yaw error angle leads to a decrease in wind turbine output power, thus affecting power generation. However, the impact of the yaw error angle on power generation is smaller at lower wind speeds. Therefore, the power generation using the optimized yaw start-up wind control strategy is only slightly lower than that using the original strategy.

[0178] Finally, wind direction and speed data recorded at a time resolution of 1 second for four consecutive months were randomly extracted from the SCADA system of the wind turbine. The two yaw start-up control strategies were compared from a long-term perspective, and the results are shown in Table 3. As can be seen from the table, after optimizing the yaw start-up wind control strategy, the number of yaws decreased significantly by 82.9%, while the power generation decreased by only 3.5%. Adopting the optimized yaw start-up control strategy can significantly reduce the number of yaws while basically ensuring the power generation of the wind turbine, thereby effectively improving the economic efficiency of the wind turbine throughout its entire life cycle.

[0179] Table 3 Comparison of power generation and yaw rate in long-term simulations

[0180]

[0181] The working principle of this embodiment is as follows:

[0182] This invention extracts operational data features and determines the wind speed segmentation scheme for the yaw system startup wind-adjustment strategy based on the actual operating characteristics of individual wind turbine units. It employs a multi-objective optimization method to find the optimal startup wind-adjustment control parameters within each wind speed segment, thereby effectively reducing the number of yaws by the wind turbine while ensuring its power generation efficiency and improving the availability and economy of the unit operation. An objective wind speed interval division standard is proposed, which is easy to implement and has the potential for widespread application. Wind speed segmentation is based on the proposed comprehensive yaw feature, which considers both the impact of the yaw error angle measured during actual wind turbine operation on the power generation efficiency and the stability of wind resources; the basis for wind speed segmentation is scientific and comprehensive. When searching for the optimal startup wind-adjustment control parameters within each wind speed segment using multi-objective optimization, the accuracy of one of the optimization objectives, power generation W, depends on the accuracy of solving the two constant coefficients in the formula for calculating the instantaneous power P of the wind turbine. Based on extensive analysis, this invention proposes a constant coefficient solution method for parameter segmentation identification according to wind speed intervals, improving the accuracy of power generation optimization target calculation. Based on the NSGA-II algorithm and the TOPSIS method, a complete multi-objective decision-making process for wind turbine yaw system startup is presented. Compared with optimization methods that transform multi-objective problems into single-objective problems, this method can better balance the two objectives of effectively reducing yaw frequency and ensuring wind turbine power generation efficiency according to specific circumstances. This invention can also be extended to multi-objective decision optimization of other complex electromechanical control systems.

[0183] The above description is one implementation method provided in conjunction with specific content, and does not imply that the specific implementation of this application is limited to these descriptions. Any methods or structures that are similar to or identical to those of this application, or any technical deductions or substitutions made based on the concept of this application, should be considered within the scope of protection of this application.

Claims

1. An optimization method for wind control strategy during the start-up of a wind turbine yaw system, characterized in that, When the wind speed is below the rated wind speed, the yaw system of the wind turbine is activated with a wind control strategy, including the following steps: S1: Read the target wind turbine's stable operation status from the SCADA system. A dataset is constructed by recording historical operational data annually with a first preset time resolution. ; and recently A dataset is constructed by recording historical running data for one month at a second preset time resolution. ; S2: Cut the target wind turbine into the wind speed range. and rated wind speed wind speed range Equal-distance discretization Sub-intervals , ... ; S3: Use Based on the extracted comprehensive yaw characteristics of the wind speed range, the data is used to determine... Wind speed segmentation scheme for yaw system to activate wind countermeasure strategy within wind speed range; S4: If the wind speed ranges that require independent setting of the yaw system to initiate the wind countermeasure strategy obtained in step S3 are total... One, based on the number of yawing times Minimize and generate electricity Maximize as the objective, using the dataset A multi-objective optimization algorithm was used to determine the yaw error angle threshold for each wind speed range. Delay time threshold Optimization is performed to obtain the Pareto front solution set of the yaw control parameters; S5: Determine the importance weights of the two optimization objectives, namely, the number of yaws and the amount of power generated, and use the TOPSIS method of multi-attribute decision analysis to obtain the optimal solution of the yaw start control parameters from the Pareto front solution set; S6: Set the yaw start control strategy for each interval to: when the first interval... Within a wind speed range where the yaw error angle exceeds a threshold, the yaw system requires independent configuration to initiate the wind countermeasure strategy. And the duration exceeds the delay time threshold. At that time, start the wind stabilization; In step S3, the yaw composite feature is calculated and synthesized from the "yaw error-power" correlation feature and the wind speed turbulence feature. The method for extracting the "yaw error-power" correlation feature of the wind speed sub-interval includes the following steps: A1: Calculate the wind speed range The benchmark power generation : in, For dataset Sample points in; It is a dataset The wind speed is within the current wind speed sub-interval and the absolute value of the yaw error angle parameter is within... A subset consisting of sample points within the range; For sample points The output active power parameters; For subset The number of sample points in the middle; A2: Calculate wind speed range Deviation in power generation : in, It is a dataset The wind speed parameters are within the current wind speed sub-interval and the absolute value of the yaw error angle parameter is within... A subset consisting of sample points within the range, where ; For subset The number of sample points in the middle; A3: Calculate the "yaw error - power" correlation characteristics for this wind speed range. : in, It is the power reference constant. In the interval Take the value from; The calculation method for the comprehensive yaw characteristic, which is derived from the correlation characteristics of "yaw error-power" and wind speed turbulence characteristics of each wind speed sub-interval, is as follows: in, Characterized by wind speed turbulence. That is, the overall characteristics of yaw. The weights for the "yaw error-power" related features are: In the interval Take the value from; In step S3, based on the comprehensive yaw characteristics of each wind speed sub-interval, determine The method for initiating a yaw system's counter-wind strategy within a wind speed range, based on a segmented wind speed scheme, includes the following steps: B1: Order Calculate the first Yaw characteristics of each wind speed sub-range ; B2: Order , will the The wind speed sub-interval is used as the first Each wind speed range requires independent configuration of the yaw system to initiate the wind countermeasure strategy, and the comprehensive yaw characteristics of that wind speed range must be considered. As the first The comprehensive yaw characteristics of each wind speed range requiring independent yaw system configuration and activation of wind countermeasures. ; B3: Order Calculate the first Yaw characteristics of each wind speed sub-range ; B4: If If established, then the first The wind speed sub-intervals are merged into the first... Each wind speed segment requires independent configuration of the yaw system to initiate the wind countermeasure strategy; the intervals are then merged and recalculated. ; Otherwise , will the The wind speed sub-interval is used as the first A wind speed segment that requires independent setting of the yaw system to initiate the wind countermeasure strategy, while also As a new ; B5: Determine the first If the wind speed sub-interval is the last discrete wind speed sub-interval, then the wind speed segmentation ends, and we obtain... If the yaw system is set within the wind speed range to initiate the wind countermeasure strategy, then proceed to step B3.

2. The method for optimizing the wind control strategy during the start-up of a wind turbine yaw system according to claim 1, characterized in that, In step S5: Optimization process objectives The calculation formula is: ,in The number of times the yaw start strategy is satisfied in the analyzed operational data; Optimization process objectives The calculation formula is: ,in, The instantaneous power of the wind turbine. For data Data recording interval, The range of values ​​is ; Instantaneous power of wind turbine The calculation formula is: in, The yaw error angle of the wind turbine. The wind turbine rotation speed, These are coefficients related to wind turbine diameter, air density, and wind turbine utilization factor. This is a compensation coefficient related to error and generator conversion efficiency.

3. The method for optimizing the wind control strategy during the start-up of a wind turbine yaw system according to claim 2, characterized in that, Calculate instantaneous power parameters , Before using the power calculation formula, you need to use the dataset. The yaw error angle, rotor speed, and output active power parameters recorded in the data are solved using the least squares method, including the following steps: C1: Extract medium wind speed Sample points within the range, Using the above sample points, the least squares method is employed to solve for the wind speed range. , , respectively denoted as , ; C2: Extract medium wind speed Using the sample points within the specified range, the least squares method is employed to solve for the wind speed within this range. , , respectively denoted as , ; C3： , The results are obtained by calculating using the formulas above.

4. The method for optimizing the wind control strategy during the start-up of a wind turbine yaw system according to claim 1, characterized in that, In step S5, the multi-objective optimization algorithm used is the second-generation non-dominated sorting genetic algorithm, namely the NSGA-II algorithm, which sets the yaw error angle threshold for each wind speed range. and delay time threshold Optimize The method includes the following steps: D1: Set the yaw error angle threshold to be optimized. The reasonable range of values ​​is limited to Delay time threshold The reasonable range is limited to ; D2: According to the NSGA-II algorithm, the optimal individual coefficient is... ,by Create a variable to be optimized for population size. The initial population; among which, , ; D3: Optimize the objective according to the NSGA-II algorithm and yaw start-up wind strategy. , The calculation method assesses the fitness of individuals in the population, performs non-dominant ranking of the population based on the dominance relationship between individuals, and divides individuals into different ranks. D4: Calculate the crowding distance of individuals in each level according to the NSGA-II algorithm, and select the next generation of individuals based on the non-dominated ranking and crowding distance; D5: Cross rate value is... The mutation rate takes the value of Crossover and mutation operations are applied to selected individuals to generate the next generation of the population; among them, , ; D6: If the maximum number of generations has been reached... If the current population is the Pareto front solution set for the yaw control parameters, then output the solution set of the current population; otherwise, go to step D3. .

5. The method for optimizing the wind control strategy during the start-up of a wind turbine yaw system according to claim 4, characterized in that, In step S5, the method of determining the optimal solution from the Pareto solution set using the TOPSIS method includes the following steps: E1: Positively process the yaw rate index for each scheme: , in, For the Pareto solution set, the first The yaw rate index for each solution; This is the positiveized yaw rate indicator. The number of solutions in the Pareto front solution set is known from the principle of the NSGA-II algorithm. ; E2: Construct standardized evaluation index vectors for each solution in the solution set. ,in, : E3: Calculate the evaluation index vector of the ideal solution. Evaluation index vector of negative ideal solution : E4: Calculate the distance from each solution to the ideal solution and the negative ideal solution. , : in, , These are the importance weights for power generation indicators and the number of yaws, respectively. The range of values ​​is and ; E5: Calculate the score for each solution. : E6: Rank the solutions according to their scores and select the solution with the highest score as the optimal solution.

6. The method for optimizing the wind control strategy during the start-up of a wind turbine yaw system according to claim 1, characterized in that, In step S2, the target wind turbine's cut-in wind speed is... The range of values ​​is Rated wind speed The range of values ​​is ; Number of discrete wind speed ranges It is a positive integer, and its value range is 1. .

Images