Yaw control method based on data driving and multi-objective optimization
By dividing the wind speed ranges in mountain wind farms and optimizing the yaw control parameters, the problems of frequent yaw movements and insufficient control accuracy were solved, achieving efficient operation and improved economic benefits of wind turbines, and adaptive adjustment to complex wind conditions.
Patent Information
- Application Number
- CN202511112184.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-08-08
- Publication Date
- 2025-10-17
AI Technical Summary
Traditional yaw control strategies in mountain wind farms have problems such as frequent yaw movements, insufficient control accuracy, imbalanced multi-objective optimization, and lack of differentiated optimization, resulting in high risk of equipment failure, low power generation efficiency, and poor economic benefits.
A data-driven and multi-objective optimization method is adopted to divide the wind speed range and set differentiated yaw control parameters. The objective function of minimizing comprehensive economic benefits and the number of yaws is constructed. The genetic algorithm and NSGA-II algorithm are used to optimize the yaw control parameters. Sensitivity analysis and simulation verification are combined to achieve adaptive adjustment.
It reduces the mechanical wear of the yaw system, improves the power generation efficiency and economic benefits of the wind turbine, enhances the operating stability and response speed of the system, and adapts to complex and changeable mountain wind conditions.
Smart Images
Figure CN120798656A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of wind turbine control, and particularly to a yaw control method based on data driving and multi-objective optimization. BACKGROUND
[0002] With the wide application of wind energy worldwide, wind farms are gradually expanding from plain areas to complex terrain mountainous areas. Due to the undulating terrain, drastic changes in wind speed and direction, and large fluctuations in turbulence intensity, yaw control of wind turbines in mountainous areas faces many challenges. The traditional yaw control strategy has obvious shortcomings in such a complex environment:
[0003] Frequent yaw action: the complex wind conditions in mountainous areas and the variable wind direction result in excessive yaw action of the wind turbine. Frequent yaw increases the risk of equipment failure, such as gear breakage, yaw bearing wear, and failure of other auxiliary equipment, which seriously affects the reliable operation of the wind turbine.
[0004] Insufficient yaw control accuracy: the traditional strategy is difficult to accurately track the wind direction, resulting in insufficient yaw control accuracy, which reduces the wind energy capture efficiency and affects the power generation efficiency of the wind turbine.
[0005] Multi-objective optimization imbalance: most existing methods simplify the complex multi-objective optimization problem into a single-objective optimization problem, which cannot effectively balance the reduction of yaw frequency and the guarantee of economic benefits.
[0006] Lack of differentiated optimization: existing technologies usually set uniform yaw control parameters and fail to optimize different wind speed intervals and complex terrain requirements, making it difficult to fully utilize the performance of the wind turbine.
[0007] Therefore, the above technical problems need to be solved. SUMMARY
[0008] In order to overcome the shortcomings of the prior art, the present application provides a yaw control method based on data driving and multi-objective optimization.
[0009] In order to solve the above technical problems, the basic technical scheme of the present application is as follows:
[0010] A yaw control method based on data driving and multi-objective optimization, comprising the following steps:
[0011] Wind speed interval division step: according to the wind resource characteristics, yaw behavior, wind direction distribution characteristics, wind speed frequency distribution, turbulence intensity in different wind speed intervals, and the influence of yaw deviation angle on output power in different wind speed intervals, the wind speed range below the rated wind speed is divided into low-speed stable interval 3-6m / s, medium-speed optimization interval 6-8m / s and high-speed balance interval 8-11m / s;
[0012] The multi-objective optimization model construction step comprises:
[0013] A comprehensive economic benefit maximization objective function is constructed, and the power generation of the wind turbine, the energy consumption of the yaw motor and the converted power due to the fatigue life loss of the yaw bearing are comprehensively considered, wherein the power generation of the wind turbine is calculated by the formula F1=P△t, P=aω 3 cos 3 θ+b, the parameters a and b are identified by using a genetic algorithm; the power consumed by the yaw motor is calculated by the formula F2=18T y ; the converted power due to the fatigue life loss of the yaw bearing is obtained by using an interpolation method according to a pre-calculated data table; and the comprehensive economic benefit is calculated by the formula F=F1-F2-F3.
[0014] A yaw frequency minimization objective function is constructed, and the yaw frequency is calculated by the formula M=T a ;
[0015] The optimization solving step comprises: performing evolutionary search on the yaw control parameters including the yaw deviation angle and the delay time by using an NSGA-II algorithm, and satisfying the constraint condition 0≤T i ≤300, wherein i=1, 2, 3; evaluating and deciding the Pareto front solution by using a sensitivity analysis method, introducing a "sensitivity" to measure the sensitivity degree of the average variation rate of a certain objective function value with respect to a unit function value, introducing a bias degree to represent the bias degree of the non-inferior solution with respect to different optimization objectives, and determining the optimal control parameters.
[0016] The simulation verification step comprises: constructing a yaw system simulation model in MATLAB software, selecting wind speed and wind direction data for yaw control parameter optimization, updating the comprehensive economic benefit, the yaw frequency and the yaw deviation angle according to the corresponding formula during the optimization process, and the update formula of the yaw deviation angle is θ1=θ1△tV y ; and verifying the effectiveness of the optimization strategy by comparing the yaw frequency, the yaw time, the power generation of the unit, the power consumed by the yaw motor, the converted power due to the bearing life loss and the comprehensive economic benefit before and after the optimization.
[0017] Preferably, in the wind speed interval division, a relatively large yaw deviation threshold and a relatively long delay time are selected in the low-speed stable interval, a relatively small yaw deviation threshold and a relatively short delay time are set in the medium-speed optimization interval, and a relatively moderate yaw deviation threshold and delay time are adopted in the high-speed balance interval.
[0018] Preferably, in the multi-objective optimization model construction step, the genetic algorithm identifies the parameters a and b by using the actually measured wind speed, wind direction, rotor speed and the like for iterative calculation.
[0019] Preferably, in the optimization solving step, the NSGA-II algorithm sets the population size, the number of iterations and other parameters, and updates the population through fast non-dominated sorting and crowded distance comparison operations.
[0020] Preferably, in the simulation verification step, wind speed and wind direction data of different time periods are selected for simulation analysis, and if the optimization effect does not reach the expectation, the wind speed interval division, optimization algorithm parameters or objective function weight are adjusted, and optimization and simulation analysis are performed again.
[0021] The beneficial effects of the present application are:
[0022] The technical scheme of the present application precisely divides the low-speed stable, medium-speed optimization and high-speed balance wind speed intervals by in-depth analysis of the wind resource characteristics and yaw behavior of the mountain wind farm, and differentiates the yaw control parameters for each interval. In the low wind speed interval, a larger yaw deviation threshold and a longer delay time effectively avoid unnecessary yaw action. Taking actual simulation data as an example, after optimization, the yaw frequency is significantly reduced in the low wind speed range of 3-6 m / s, which reduces a certain proportion compared with before optimization, which greatly reduces the mechanical wear of the yaw system and prolongs the service life of key components such as yaw bearings and gears.
[0023] The optimized yaw control strategy realizes the goal of efficient capture of wind energy in each wind speed interval. On the one hand, in the medium and high speed intervals, reasonable yaw control parameters ensure that the wind turbine can quickly and accurately track the wind direction, improving the power generation efficiency; on the other hand, by reducing the yaw frequency, the energy consumption of the yaw system is reduced, and the loss of yaw bearing fatigue life is reduced. From the simulation comparison results, the comprehensive economic benefit of the wind turbine after optimization is significantly improved, effectively reducing the operation and maintenance cost and improving the economic benefit of the wind turbine throughout its life cycle.
[0024] The present application adopts data-driven and multi-objective optimization algorithm, so that the yaw system can adjust the yaw control parameters flexibly according to the dynamic changes of wind speed, wind direction and turbulence intensity. Whether it is a sudden change of wind direction or a fluctuation of turbulence intensity, the yaw system can respond quickly to ensure that the wind turbine is always in the best windward state. This adaptive adjustment capability greatly improves the operation stability and response speed of the entire system, making the wind turbine better adapt to complex and variable mountain wind conditions. BRIEF DESCRIPTION OF DRAWINGS
[0025] Figure 1 is the working principle diagram of the yaw system of the present application;
[0026] Figure 2 is the wind direction rose of the present application;
[0027] Figure 3 Wind speed frequency distribution diagram of the present application;
[0028] Figure 4 NSGA-II algorithm optimization flow chart of the present application;
[0029] Figure 5 Pareto front diagram of NSGA-II algorithm optimization of the present application. DETAILED DESCRIPTION
[0030] The technical solutions in the embodiments of the present application will be described in detail below with reference to the accompanying drawings. Figure 1 to the accompanying drawings Figure 5 The technical solutions in the embodiments of the present application will be described in detail below with reference to the accompanying drawings.
[0031] A yaw control method based on data driving and multi-objective optimization, comprising the following steps:
[0032] Wind speed interval division step: according to the wind resource characteristics, yaw behavior, wind direction distribution characteristics, wind speed frequency distribution, turbulence intensity in different wind speed intervals, and the influence of yaw deviation angle on output power in different wind speed intervals, the wind speed range below the rated wind speed is divided into low-speed stable interval 3-6 m / s, medium-speed optimization interval 6-8 m / s and high-speed balance interval 8-11 m / s; in the wind speed interval division, a relatively large yaw deviation threshold and a longer delay time are selected for the low-speed stable interval, a relatively small yaw deviation threshold and a shorter delay time are set for the medium-speed optimization interval, and a relatively moderate yaw deviation threshold and delay time are used for the high-speed balance interval.
[0033] In the low-speed stable interval, the wind speed is relatively low, the turbulence intensity is high, and the yaw adjustment has limited effect on the improvement of power generation efficiency. Therefore, a relatively large yaw deviation threshold and a longer delay time are selected to reduce unnecessary yaw times, reduce system wear and tear and improve the stability of unit operation; in the medium-speed optimization interval, the turbulence intensity gradually decreases, and the influence of yaw deviation angle on power generation efficiency is significantly enhanced. At this time, a relatively small yaw deviation threshold and a shorter delay time are set to ensure that the wind turbine can quickly and accurately track the wind direction and maximize the capture efficiency of wind energy. In the high-speed balance interval, as the wind speed further increases, the turbulence intensity gradually tends to be stable, and the influence of yaw deviation angle on power generation efficiency is relatively small. A relatively moderate yaw deviation threshold and delay time are used to ensure high power generation efficiency while avoiding the aggravation of system burden caused by frequent yaw.
[0034] Multi-objective optimization model construction step:
[0035] A comprehensive economic benefit maximization objective function is constructed, and the power generation of the wind turbine, the energy consumption of the yaw motor and the converted power due to the fatigue life loss of the yaw bearing are comprehensively considered, wherein the power generation of the wind turbine is calculated by formula F1=P△t, the output power P=aω3cos3θ+b, wherein P is the output power of the wind turbine, △t is the sampling period of SCADA data, due to the difficulty in accurately measuring the related variables in practice, the genetic algorithm is used to identify the two parameters. Each unit is equipped with 6 yaw motors with a power of 3kW
[0036] The parameters a and b are identified by using the genetic algorithm; the power consumed by the yaw motor is represented by F2=18Ty; Ty is the total time of yaw action, and the converted power due to the fatigue life loss of the yaw bearing is obtained by using the interpolation method according to the pre-calculated data table; the comprehensive economic benefit is calculated by formula F=F1-F2-F3; in the multi-objective optimization model construction step, when the genetic algorithm identifies the parameters a and b, the measured wind speed, wind direction, rotor speed and other data are used for iterative calculation.
[0037] A yaw frequency minimization objective function is constructed, and the yaw frequency is calculated by formula M=T a , T a is the number of time points that meet the initial yaw condition and are in the non-yaw state;
[0038] Optimization solving step: the NSGA-II algorithm is used to perform evolutionary search on the yaw control parameters, including the yaw deviation angle and the delay time, and the constraint condition 0≤T i ≤300, wherein i=1, 2, 3, the sensitivity analysis method is used to evaluate and decide the Pareto front solution, the average variability between each non-boundary solution and its adjacent solution is calculated, the "sensitivity" is introduced to measure the sensitivity of the average variability of a certain objective function value to the unit function value, the bias degree is introduced to represent the bias degree of the non-inferior solution to different optimization objectives, and the optimal control parameter is determined. In the optimization solving step, the NSGA-II algorithm sets the population size, the number of iterations and other parameters, and updates the population through fast non-dominated sorting and crowded distance comparison operation.
[0039] Simulation verification step: a yaw system simulation model is constructed in MATLAB software, and wind speed and wind direction data are selected for yaw control parameter optimization, in the optimization process, the comprehensive economic benefit, the yaw frequency and the yaw deviation angle are updated according to the corresponding formula, the update formula of the yaw deviation angle is θ1=θ1△tV y , θ1 is the updated yaw deviation angle, θ1 is the updated yaw deviation angle, V yFor yaw angle velocity, the effectiveness of the optimization strategy is verified by comparing the number of yawing, yawing time, unit power generation, power consumption of yawing motor, bearing life loss equivalent power and comprehensive economic benefit before and after optimization. In the simulation verification step, the wind speed and wind direction data of different time periods are selected for simulation analysis. If the optimization effect does not meet the expectation, adjust the wind speed interval division, optimization algorithm parameters or objective function weight, and optimize and simulate again. The NSGA-II algorithm is used to optimize the yaw control parameters, the sensitivity analysis method is used to determine the final optimization result, and the weight values of the two optimization objectives are set according to the actual demand. The effectiveness of the optimization strategy is verified by comparing the number of yawing, yawing time, unit power generation, power consumption of yawing motor, bearing life loss equivalent power and comprehensive economic benefit before and after optimization.
[0040] The specific implementation is as follows:
[0041] Wind speed interval division implementation: Collect historical wind speed, wind direction, turbulence intensity and other data of mountain wind farm, combine with wind turbine operation data, analyze wind resource characteristics and yaw behavior in different wind speed intervals. According to the analysis results, the wind speed range below rated wind speed is divided into low speed stable interval, medium speed optimization interval and high speed balance interval according to the above rules.
[0042] Multi-objective optimization model construction and solution implementation: According to the parameters of wind turbine, initialize each parameter in multi-objective optimization model. Use genetic algorithm to identify parameters a and b in output power model, substitute actual measured wind speed, wind direction, rotor speed and other data into genetic algorithm for iterative calculation to get appropriate a and b values. Use NSGA-II algorithm to optimize yaw control parameters, set population size, iteration number and other parameters of the algorithm, and perform evolutionary search under the condition of meeting the constraint conditions. In each iteration process, calculate the values of two objective functions (comprehensive economic benefit and yawing frequency) of each individual, update the population through fast non-dominated sorting and crowded distance comparison operation. Use sensitivity analysis method to evaluate and decide the final Pareto frontier solution, set the weight values of comprehensive economic benefit and yawing frequency according to actual demand, for example, set the weight of comprehensive economic benefit as 0.75 and the weight of yawing frequency as 0.25, select the optimal yaw control parameter combination.
[0043] Simulation analysis implementation: In MATLAB software, the simulation model of yaw system is built, and the optimized yaw control parameters are substituted into the model. The wind speed and wind direction data of different time periods, such as one day, two days or one month, are selected for simulation analysis. In the simulation process, the comprehensive economic benefit, yaw frequency and yaw deviation angle are calculated according to the corresponding update formula. The indicators before and after optimization, such as yaw frequency, yaw time, unit power generation, power consumption of yaw motor, bearing life loss equivalent power and comprehensive economic benefit, are compared to evaluate the effectiveness of the optimization strategy. If the optimization effect does not meet the expectation, the wind speed interval division, optimization algorithm parameters or objective function weight can be adjusted for optimization and simulation analysis again until a satisfactory result is obtained.
[0044] According to the disclosure and teaching of the above description, those skilled in the art of the present application can also make changes and modifications to the above embodiments. Therefore, the present application is not limited to the specific embodiments disclosed and described above, and some modifications and changes of the present application should fall within the protection scope of the claims of the present application. In addition, although some specific terms are used in the specification, these terms are only for convenience of description and do not constitute any limitation on the present application.
Claims
1. A yaw control method based on data-driven and multi-objective optimization, characterized in that: The steps include: Wind speed range division steps: Based on the wind resource characteristics, yaw behavior, wind direction distribution characteristics, wind speed frequency distribution, turbulence intensity in different wind speed ranges, and the impact of yaw deviation angle on output power in different wind speed ranges of mountain wind farms, the wind speed range below the rated wind speed is divided into a low-speed stable range of 3-6m / s, a medium-speed optimization range of 6-8m / s, and a high-speed balance range of 8-11m / s; Steps to build a multi-objective optimization model: The objective function of maximizing comprehensive economic benefits is constructed, which comprehensively considers the power generation of the wind turbine, the energy consumption of driving the yaw motor, and the converted power loss due to the fatigue life of the yaw bearing. The power generation of the wind turbine is calculated by the formula F1 = P△t, P = aω 3 cos 3 θ+b, using genetic algorithm to identify parameters a and b; the power consumed by the yaw motor is calculated by the formula F2=18T y Calculation: The converted power of the yaw bearing fatigue life loss is obtained by interpolation based on the pre-calculated data table; the comprehensive economic benefits are calculated using the formula F = F1-F2-F3; Construct the objective function of minimizing the number of yaws, and the number of yaws is calculated by the formula M=T a calculate; Optimization solution steps: Use NSGA-II algorithm to perform evolutionary search on the yaw control parameters, which include yaw deviation angle and delay time, and satisfy the constraint condition 0≤T i ≤300, where i = 1, 2, 3. The sensitivity analysis method is used to evaluate and make decisions on the Pareto frontier solutions. By calculating the average rate of change between each non-boundary solution and its adjacent solutions, "sensitivity" is introduced to measure the sensitivity of the average rate of change of a certain objective function value relative to the unit function value. The bias degree is introduced to indicate the bias degree of the non-inferior solution relative to different optimization objectives, and the optimal control parameters are determined. Simulation verification steps: Build a yaw system simulation model in MATLAB software, select wind speed and direction data to optimize yaw control parameters, and update the comprehensive economic benefits, yaw times and yaw deviation angle according to the corresponding formula during the optimization process. The update formula of yaw deviation angle is θ1=θ1△tV y ,The effectiveness of the optimization strategy was verified by comparing the yaw times, yaw time, unit power generation, power consumption by the yaw motor, power consumption converted from bearing life loss, and ,comprehensive economic benefits before and after optimization.
2. The yaw control method based on data-driven and multi-objective optimization according to claim 1, characterized in that: In the wind speed interval division, a relatively large yaw deviation threshold and a long delay time are selected for the low-speed stable interval, a relatively small yaw deviation threshold and a short delay time are set for the medium-speed optimization interval, and a relatively moderate yaw deviation threshold and delay time are adopted for the high-speed balance interval.
3. The yaw control method based on data-driven and multi-objective optimization according to claim 1, characterized in that: In the multi-objective optimization model construction step, when the genetic algorithm identifies parameters a and b, it uses the actual measured wind speed, wind direction, rotor speed and other data for iterative calculation.
4. The yaw control method based on data-driven and multi-objective optimization according to claim 1, characterized in that: In the optimization solution step, the NSGA-II algorithm sets parameters such as population size and number of iterations, and updates the population through fast non-dominated sorting and crowding distance comparison operations.
5. The yaw control method based on data-driven and multi-objective optimization according to claim 1, characterized in that: In the simulation verification step, wind speed and wind direction data of different time periods are selected for simulation analysis. If the optimization effect does not meet expectations, the wind speed interval division, optimization algorithm parameters or objective function weights are adjusted, and optimization and simulation analysis are performed again.