A wind farm multi-objective yaw control method considering comprehensive benefits of power load

Abstract

The application discloses a kind of wind farm multi-objective yaw control methods considering power load comprehensive benefits, belong to wind power generation technical field.The wind farm whole tail flow modeling of yaw state is considered by analyzing tail flow model, the inflow data of each wind turbine in wind farm is determined, and the total power and average life of wind farm are quickly and accurately predicted by combining wind turbine power, thrust coefficient curve and S-N curve of transition section pipe node.Single wind turbine power optimization and wind turbine fatigue life optimization with power constraint are integrated to establish a two-stage multi-objective control optimization system, and by building a Bayesian machine learning network, the optimal yaw combination of each stage is located quickly by using efficient and accurate power and life prediction data.The application can accurately and efficiently provide the optimal yaw strategy considering power load comprehensive benefits according to real-time wind speed data, while meeting the power generation demand of wind farm, the load is maximized to extend the service life of wind turbine.

Description

A Multi-Objective Yaw Control Method for Wind Farms Considering the Comprehensive Benefits of Power Load Technical Field

[0001] This invention belongs to the field of wind power generation technology, specifically relating to a multi-objective yaw control method for wind farms that considers the comprehensive benefits of power load. Background Technology

[0002] With the increasing global demand for renewable energy, wind power, as an important green energy source, has become a crucial component of energy structure transformation in various countries. As the core unit of wind power generation, the operating efficiency and economics of wind farms directly impact the overall contribution of wind power. However, in the actual operation of wind farms, the wake effect generated by the interaction between wind turbines, as well as the instability of wind speed and direction, significantly affect the power output and fatigue load of the turbines. Reasonable yaw control can significantly mitigate the wake effect, playing a vital role in optimizing the power generation efficiency of wind farms and reducing mechanical wear and fatigue damage.

[0003] Traditional wind farm yaw control methods mostly focus on optimizing a single objective, typically prioritizing maximizing power output. For example, patent application CN 110728066 A discloses a wind farm sector optimization method and system. This method combines historical data-determined power curves and model-calculated effective wind speeds to calculate the total power of the wind farm, and uses an optimization algorithm to determine the optimal yaw angle when the total power is maximized, applied to the rapid and effective optimization of wind farm sectors. While this method can improve the power generation efficiency of wind turbines to some extent, fatigue loads on turbine components gradually emerge during long-term operation, especially when the turbine is in the wake region for extended periods or when wind speed changes drastically. Over-reliance on power optimization may exacerbate fatigue damage to turbine components, thereby shortening the turbine's lifespan and increasing maintenance and replacement costs. Therefore, balancing power output and turbine fatigue loads in a wind farm has become a key issue in current wind farm control technology research. Furthermore, considering wind speed, wind direction changes, and wake effects in wind farms, more efficient fatigue life prediction models are needed to account for the interaction between the turbine and the wind farm environment, achieving more accurate fatigue life assessments. Summary of the Invention

[0004] To address the problems of existing technologies, this invention provides a multi-objective yaw control method for wind farms that considers the comprehensive benefits of power load. This method can optimize the yaw angle in real time through a two-stage multi-objective control strategy, thereby improving the power generation efficiency of the wind farm, effectively extending the service life of the wind turbine, reducing the maintenance cost of the wind farm, and thus improving the overall economic efficiency and sustainability of the wind farm.

[0005] The present invention adopts the following technical solution:

[0006] A multi-objective yaw control method for wind farms that considers the comprehensive benefits of power load includes the following steps:

[0007] Step 1: Obtain the layout information of the wind farm, the basic parameters of the wind turbines, and the measured data of wind speed and direction;

[0008] The basic parameters of the wind turbine include: rotor diameter D, hub height H, power-wind speed curve, thrust coefficient curve, and SN curve of the transition section pipe node of the wind turbine.

[0009] Step 2: Establish a local three-dimensional coordinate system with the current wind flow direction as the positive x-axis, and determine the local three-dimensional coordinate data of each wind turbine in the wind farm based on the wind farm layout information obtained in Step 1.

[0010] Step 3: Based on the measured wind speed and direction data obtained in Step 1 and the local three-dimensional coordinate data of the wind turbine obtained in Step 2, the wake of the entire wind farm considering the yaw state is modeled using the yaw analytical wake model, and a wake prediction model for the entire wind farm is established.

[0011] Step 4: Establish the relationship between the wind farm yaw control strategy and total power generation. Based on the whole-field wake prediction model established in step 3, the inflow wind speed of each wind turbine in the wind farm is determined. Combined with the power-wind speed curve, the power of each wind turbine is calculated and the total power of the wind farm is obtained by summing them up.

[0012] Step 5: Establish the relationship between wind farm yaw control strategy and mean fatigue life. Based on the whole-field wake prediction model established in step 3, the inflow wind speed and turbulence intensity of each wind turbine in the wind farm are determined. Combined with the SN curve of the pipe node of the wind turbine transition section, the fatigue life of each wind turbine is calculated, and the average value is taken to obtain the average life of the wind farm.

[0013] Step 6: Establish the relationship between the wind farm yaw control strategy and total power generation based on the data from Step 4. By combining a Bayesian optimization learning network, the first stage of yaw control optimization considering the single objective function of power generation is carried out, and the optimal yaw combination and corresponding total power are output as the baseline yaw strategy and baseline power. At the same time, the relationship between the wind farm yaw control strategy and the average fatigue life established in step 5 is used to calculate the wind farm fatigue life under the baseline yaw strategy, which is used as the baseline average life.

[0014] Step 7: Based on the relationship between the wind farm yaw control strategy and total power generation. The relationship between wind farm yaw control strategy and mean fatigue life Combining a Bayesian optimization learning network, with the lifetime extension factor relative to the baseline average lifetime as the optimization objective, and based on the actual power generation demand of the wind farm, the allowable reduction coefficient of the relative baseline power is set as a constraint, and the second stage of multi-objective yaw control optimization considering the comprehensive benefits of power load is carried out.

[0015] Step 8: Output the final optimal yaw control strategy and its corresponding total power and average fatigue life, and visualize the wake distribution of the entire field to complete the multi-objective yaw control of the wind farm that considers the comprehensive benefits of power load.

[0016] Preferably, the layout information of the wind farm in step 1 is the global three-dimensional coordinate data of each wind turbine in a global three-dimensional coordinate system with due east as the positive x-axis direction.

[0017] Preferably, in step 2, determining the local three-dimensional coordinate data of each wind turbine in the wind farm based on the layout information of the wind farm obtained in step 1 specifically involves: converting the global three-dimensional coordinate data of each wind turbine obtained in step 1 into local three-dimensional coordinate data of each wind turbine in a local three-dimensional coordinate system.

[0018] The conversion between the global 3D coordinate data and the local 3D coordinate data of the wind turbine satisfies the following relationship:

[0019] ;

[0020] Where (x, y) represents the global three-dimensional coordinate data of the wind turbine, and (x1, y1) represents the local three-dimensional coordinate data of the wind turbine. The angle between the local coordinate system and the global coordinate system in the counterclockwise direction.

[0021] Preferably, the process of establishing the whole-field wake prediction model in step 3 is as follows:

[0022] Step 3.1: Number all the fans in ascending order of their horizontal coordinates (1, …, N), where N represents the number of fans;

[0023] Step 3.2: Based on the wind turbine number, determine the inflow wind speed and turbulence intensity of each wind turbine in the wind farm in sequence. The inflow wind speed and turbulence intensity of the first wind turbine are measured wind speed data. For the i-th downstream wind turbine, its inflow conditions are obtained by superimposing the wakes of all upstream wind turbines. The inflow wind speed and turbulence intensity are superimposed using the sum of squares and linear superposition methods, respectively.

[0024] Step 3.3: Based on the inflow wind speed and turbulence intensity of each wind turbine, the inflow wind speed and turbulence intensity at any location in the downstream wake region are calculated using a yaw analytical wake model that fully considers the effects of atmospheric turbulence intensity, thrust coefficient, and yaw angle. The velocity deficit and additional turbulence exhibit Gaussian and double Gaussian distributions radially, respectively, as shown in the following equation:

[0025] ;

[0026] ;

[0027] in, For the inflow wind speed, For turbulence intensity, and For upstream undisturbed wind speed and turbulence intensity, The thrust coefficient to account for yaw effect, Where is the inflow turbulence intensity to the fan, D is the fan rotor diameter, and s is the wake diffusion distance. To account for the radial distance of the yaw wake centerline offset, K is the standard deviation of the Gaussian function, and k1 and k2 are the wake diffusion coefficients.

[0028] Preferably, the fatigue life calculation for each wind turbine in step 5 specifically includes:

[0029] Step 5.1: Based on the average wind speed and turbulence intensity of the incoming flow to the fan, and combined with the fan thrust coefficient curve, calculate the peak stress respectively. and mean stress :

[0030] ;

[0031] in, and These represent the average and peak wind speeds of the incoming airflow to the fan, respectively. Where is the pile radius, Let be the moment of inertia of the pile section, h be the height of the wind turbine hub, and A be the swept area of ​​the wind turbine. air density;

[0032] Step 5.2: Calculate the alternating stress amplitude based on the peak stress and the mean stress. To account for the influence of mean stress on the structural fatigue life, the Goodman curve was used to correct it, thus obtaining the equivalent alternating stress amplitude. :

[0033] ;

[0034] ;

[0035] in, For ultimate strength, For yield limit, To correct the mean stress;

[0036] Step 5.3: Based on the SN curve of the pipe node in the transition section of the wind turbine, according to the equivalent alternating stress amplitude... Calculate the corresponding fatigue life, i.e., the number of failure cycles N:

[0037] ;

[0038] ;

[0039] Where t is the thickness of the area where fatigue cracks may grow. For reference thickness, the value is 32 mm for tubular nodes, and k is the thickness coefficient.

[0040] Preferably, step 6, which utilizes Bayesian machine learning to optimize yaw control in the first stage considering a single objective function of power generation, specifically includes:

[0041] Step 6.1: Randomly generate a set of yaw control strategies, and use the relationship between the wind farm yaw control strategy and the total power generation established in Step 4. Obtain the corresponding total power and generate an initial training set consisting of the yaw control strategy and the corresponding total power;

[0042] Step 6.2: The training set is used to establish an approximate probability distribution between the yaw control strategy and the total power using a Gaussian process, and the expectation and variance of the power prediction probability distribution are obtained;

[0043] Step 6.3: Based on the obtained expectation and variance, establish the acquisition function, search for possible optimal yaw combinations, and use the relationship between the wind farm yaw control strategy and the total power generation established in Step 4 to obtain the corresponding total power, form new data points and incorporate them into the training set.

[0044] Step 6.4: Using the training set obtained in Step 6.3, repeatedly generate the power prediction probability distribution and search for possible optimal yaw control strategies until the optimization convergence condition is met.

[0045] Preferably, in step 7, the second stage of multi-objective yaw control optimization considering the comprehensive benefits of power load incorporates the power constraint considering the allowable reduction coefficient into the lifetime extension factor of the objective function in the form of a penalty function. Based on the relationship between the wind farm yaw control strategy and the total power generation established in step 4 and the relationship between the wind farm yaw control strategy and the average fatigue life established in step 5, a new relationship between the objective value and the yaw control strategy is established:

[0046] ;

[0047] Among them, L baseline and P baseline These are the baseline average lifetime and baseline power obtained in step 6, respectively. 0 represents the allowable power reduction factor. This is a penalty factor.

[0048] The beneficial effects of this invention are:

[0049] This invention employs the advanced Qian & Ishihara yaw analytical wake model, fully considering the influence of atmospheric turbulence, thrust coefficient, and yaw angle on the wake distribution. It accurately and rapidly describes the wake field under wind turbine yaw conditions, simulating the relationship between wind farm yaw control strategies and total power generation under real-world operating conditions. Simultaneously, it creatively proposes a novel method for rapid prediction of wind farm fatigue life. This method combines velocity and turbulence predictions based on the yaw wake model with the SN curves of the turbine transition section pipe nodes, enabling real-time prediction of the wind farm's average fatigue life. This provides strong technical support for wind farm optimization problems requiring extensive iterative calculations. Furthermore, it innovatively integrates single turbine power optimization with power-constrained turbine fatigue life optimization, establishing a two-stage multi-objective control optimization system. This system minimizes load and extends turbine service life while meeting the wind farm's power generation needs. Moreover, it fully utilizes the advantage of Bayesian optimization networks, which can rapidly converge to the global optimum with a small number of samples, accelerating the location of the optimal yaw combination at each stage of multi-objective control. The method described in this invention is applicable to multi-objective real-time yaw control of any wind farm considering the comprehensive benefits of power load. Based on measured wind speed data, it can accurately and efficiently provide the optimal yaw control strategy, effectively solving the problems in multi-objective yaw control of wind farms, such as the lack of effective fatigue life prediction methods, reasonable power and fatigue life coupling methods, and high-precision and efficient optimization algorithms. Thus, it optimizes the power generation efficiency of wind farms while effectively extending the service life of wind turbines, achieving cost reduction and efficiency improvement. Attached Figure Description

[0050] Figure 1 is a schematic diagram of the method flow of the present invention.

[0051] Figure 2 is a schematic diagram of the wind farm layout in Example 1.

[0052] In Figure 3, (a) is the power-wind speed diagram of the wind turbine selected for the wind power site in Example 1; (b) is the thrust coefficient curve of Example 1.

[0053] Figure 4(a) shows the distribution of inflow velocity in the Qian & Ishihara yaw wake model; (b) shows the distribution of turbulence intensity.

[0054] Figure 5 is a schematic diagram of the Bayesian machine learning optimization process in step 6.

[0055] Figure 6 shows the single power optimization results of the first stage of the wind farm in Example 1.

[0056] Figure 7 shows the multi-objective yaw control results for the wind farm in Example 1 under different allowable power reduction factors: (a) 0 = 96%; (b) 0 = 97%; (c) 0 = 98%; (d) 0 = 99%; (e) 0 = 100%.

[0057] Figure 8 shows the changes in actual power reduction and lifetime growth of the wind farm in Example 1 as a function of the allowable power reduction factor. Detailed Implementation

[0058] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings and specific examples:

[0059] Example 1

[0060] As shown in Figure 1, this embodiment takes a small wind farm consisting of 5 wind turbines arranged in a straight line as an example. As shown in Figure 2, the wind turbine model selected is Vestas V80 2MW. In practical applications, the time-series inflow information can be equivalent to the sum of multiple steady-state flow fields considering the average inflow within the control interval. Therefore, this embodiment focuses on a certain steady-state fixed inflow: the wind speed at the wind turbine hub height is 12 m / s, and the turbulence intensity is 4%. The specific steps are as follows:

[0061] Referring to Figures 1 to 8, a multi-objective yaw control method for wind farms considering the comprehensive benefits of power load is characterized by comprising the following steps:

[0062] Step 1: Obtain the layout information of the wind farm, the basic parameters of the wind turbines, and the measured data of wind speed and direction.

[0063] The layout information of the wind farm is the global three-dimensional coordinate data of each wind turbine in a global three-dimensional coordinate system with due east as the positive x-axis direction.

[0064] In this embodiment, the basic parameters of the wind turbine are as follows: the interval between adjacent wind turbines is 5D, the wind turbine rotor diameter D = 80 m, the hub height H = 70 m, and the power-wind speed and thrust coefficient curves are shown in Figure 3. The SN curve of the wind turbine transition section pipe node satisfies the following relationship:

[0065] ;

[0066] .

[0067] The measured wind speed and direction data are as follows: the inflow wind speed at the wind farm is 12 m / s, the turbulence intensity is 4%, and the wind direction is along the axis of the wind turbine arrangement.

[0068] Step 2: Establish a local three-dimensional coordinate system with the current wind flow direction as the positive x-axis, and determine the local three-dimensional coordinate data of each wind turbine in the wind farm based on the wind farm layout information obtained in Step 1.

[0069] The global three-dimensional coordinate data of each wind turbine obtained in step 1 is converted into local three-dimensional coordinate data of each wind turbine in a local three-dimensional coordinate system.

[0070] The conversion between the global 3D coordinate data and the local 3D coordinate data of the wind turbine satisfies the following relationship:

[0071] ;

[0072] Where (x, y) represents the global three-dimensional coordinate data of the wind turbine, and (x1, y1) represents the local three-dimensional coordinate data of the wind turbine. The angle between the local coordinate system and the global coordinate system in the counterclockwise direction.

[0073] The local three-dimensional coordinate data of the five wind turbines obtained in this embodiment are (0,0), (5D,0), (10D,0), (15D,0), and (20D,0).

[0074] Step 3: Based on the measured wind speed and direction data obtained in Step 1 and the local three-dimensional coordinate data of the wind turbine obtained in Step 2, the wake of the entire wind farm considering the yaw state is modeled using the yaw analytical wake model, and a whole-farm wake prediction model is established.

[0075] Step 3.1: Number all the fans in ascending order of their horizontal coordinates (1, …, 5).

[0076] Step 3.2: Based on the turbine number, determine the inflow velocity and turbulence intensity of each turbine in the wind farm. The inflow velocity and turbulence intensity of the first turbine are measured wind speed data, representing a undisturbed free flow. For the downstream i-th turbine, its inflow condition (u... i ,I i The result is obtained by superimposing the wakes of all upstream wind turbines, where the inflow velocity and turbulence intensity are superimposed using the sum of squares and linear superposition methods, respectively.

[0077] The sum of squares model is used to accumulate speed losses: The superposition of turbulent kinetic energy adopts a linear model: , subscript Indicates wind turbine The upstream wind turbine.

[0078] This represents the inflow velocity of the i-th fan. Represents the inflow velocity of the j-th fan. This represents the wake velocity of the j-th wind turbine at the i-th wind turbine. Represents the inflow turbulent kinetic energy of the i-th wind turbine. Represents the turbulent kinetic energy without upstream disturbance. It is the additional turbulent kinetic energy of the wake of the j-th wind turbine at the i-th wind turbine.

[0079] Step 3.3: Based on the inflow velocity and turbulence intensity of each wind turbine, the Qian & Ishihara yaw analytical wake model, which fully considers the effects of atmospheric turbulence intensity, thrust coefficient, and yaw angle, is used to calculate the inflow velocity and turbulence intensity at any location in the downstream wake region. The velocity deficit and additional turbulence exhibit Gaussian and double Gaussian distributions radially, respectively, as shown in Figure 4 and the following equation:

[0080] ;

[0081] ;

[0082] in, For the inflow wind speed, For turbulence intensity, and For upstream undisturbed wind speed and turbulence intensity, = 12 m / s, = 4%, The thrust coefficient to account for yaw effect, Where is the inflow turbulence intensity to the fan, D is the fan rotor diameter, and s is the wake diffusion distance. To account for the radial distance of the yaw wake centerline offset, K is the standard deviation of the Gaussian function, and k1 and k2 are the wake diffusion coefficients.

[0083] Step 4: Establish the relationship between the wind farm yaw control strategy and total power generation. Based on the whole field wake prediction model established in step 3, the effective inflow wind speed of each wind turbine in the wind farm is determined. Combined with the power-wind speed curve (Figure 3(a)), the power of each wind turbine is calculated and the total power of the wind farm is obtained by summing them.

[0084] Step 5: Establish the relationship between wind farm yaw control strategy and mean fatigue life. Based on the whole-field wake prediction model established in step 3, the inflow effective wind speed and turbulence intensity of each wind turbine in the wind farm are determined. Combined with the SN curve of the wind turbine transition section pipe node, the fatigue life of each wind turbine is calculated, and the average value is taken to obtain the average life of the wind farm.

[0085] The fatigue life calculation for each wind turbine specifically includes:

[0086] Step 5.1: Based on the average wind speed and turbulence intensity of the incoming flow to the fan, and combined with the fan thrust coefficient curve (Figure 3(b)), calculate the peak stress respectively. and mean stress :

[0087] ;

[0088] in, and These represent the average and peak wind speeds of the incoming airflow to the fan, respectively. Where is the pile radius, Let be the moment of inertia of the pile section, h be the height of the wind turbine hub, and A be the swept area of ​​the wind turbine. This refers to air density.

[0089] Step 5.2: Calculate the alternating stress amplitude based on the peak stress and the mean stress. To account for the influence of mean stress on the structural fatigue life, the Goodman curve (Goodman fatigue limit line) is used for correction to obtain the equivalent alternating stress amplitude. :

[0090] ;

[0091] ;

[0092] in, For ultimate strength, For yield limit, To correct the mean stress.

[0093] Step 5.3: Based on the SN curve of the pipe node in the transition section of the wind turbine, according to the equivalent alternating stress amplitude... Calculate the corresponding fatigue life, i.e., the number of failure cycles N:

[0094] ;

[0095] ;

[0096] Where t is the thickness of the area where fatigue cracks may grow. For reference thickness, for tubular nodes, =32 mm, k is the thickness coefficient.

[0097] The average fatigue life of all wind turbines in the wind farm is taken as the average life L of the wind farm, and finally, a yaw control strategy for the wind farm is established. Relationship with mean fatigue life L :

[0098] ;

[0099] This represents the fatigue life of the i-th fan.

[0100] Step 6: Establish the relationship between the wind farm yaw control strategy and total power generation based on the data from Step 4. By combining a Bayesian optimization learning network, the first stage of yaw control optimization is carried out, which considers the single objective function of power generation. The optimal yaw combination and the corresponding total power are output as the baseline yaw strategy and baseline power.

[0101] Simultaneously, the relationship between the wind farm yaw control strategy and the mean fatigue life established in step 5 is utilized. The fatigue life of the wind farm under the baseline yaw strategy is calculated and used as the baseline average life.

[0102] The first stage of yaw control optimization using Bayesian machine learning, considering a single objective function of power generation, specifically includes:

[0103] Step 6.1: Randomly generate a set of yaw control strategies, and use the relationship between the wind farm yaw control strategy and the total power generation established in Step 4. Obtain the corresponding total power, and generate an initial training set D consisting of the yaw control strategy and the corresponding total power. 1 The superscript 1 indicates the initial training set used for the first optimization iteration.

[0104] Step 6.2: For the training set D, a Gaussian process is used to establish an approximate probability distribution between the yaw control strategy and the total power, and the power prediction probability distribution P( D n The expected value and variance N( , ), where the superscript n represents the training set used for the nth optimization iteration.

[0105] Step 6.3: Based on the obtained expectation and variance, establish the acquisition function. In this embodiment, the expected improvement (EI) function is used to search for possible optimal yaw combinations, and the relationship between the wind farm yaw control strategy and the total power generation established in Step 4 is utilized. The corresponding total power is obtained and used to form new data points, which are then incorporated into the training set D.

[0106] Step 6.4: Using the training set D obtained in Step 6.3, repeatedly generate the power prediction probability distribution and search for possible optimal yaw control strategies until the optimization convergence condition is met.

[0107] This embodiment ultimately yields the optimal yaw combination shown in Figure 6, while simultaneously outputting the corresponding total power and fatigue life as the baseline power P. baseline and baseline mean lifetime Lbaseline .

[0108] Step 7: Based on the single-power optimization results of the first stage, and according to the relationship between the wind farm yaw control strategy and the total power generation, and the relationship between the wind farm yaw control strategy and the average fatigue life, combine a Bayesian optimization learning network, with the average fatigue life L relative to the baseline. baseline Life extension factor To optimize the target, different allowable reduction factors for relative baseline power are set based on the actual power generation needs of the wind farm. As a constraint, a second-stage multi-objective yaw control optimization considering the comprehensive benefits of power load is carried out.

[0109] maximize

[0110] Meet the conditions

[0111] Power reduction constraint:

[0112] The second stage considers multi-objective yaw control optimization that takes into account the comprehensive benefits of power load, and incorporates the power constraint into the lifetime extension factor in the form of a penalty function to establish a new objective function. This is used to replace the single power objective function in the Bayesian learning framework established in step 6. Based on the relationship between the wind farm yaw control strategy and total power generation established in step 4 and the relationship between the wind farm yaw control strategy and average fatigue life established in step 5, a new relationship between the objective value and the yaw control strategy is established:

[0113] .

[0114] Among them, L baseline and P baseline These are the baseline average lifetime and baseline power obtained in step 6, respectively. 0 represents the allowable power reduction factor. This is a penalty factor.

[0115] Step 8: Output the final optimal yaw control strategy and its corresponding total power and average fatigue life, and visualize the wake distribution of the entire field to complete the multi-objective yaw control of the wind farm that considers the comprehensive benefits of power load.

[0116] Figure 7 shows the multi-objective yaw control results for a selected wind farm under different allowable power reduction factors, represented by a velocity distribution contour map. 0 = 100% represents the case without baseline power reduction, corresponding to the single power target optimization in the first stage. As shown in the figure, under different allowable power reduction coefficients, the yaw angle distribution shows a gradual decreasing trend along the flow direction. The first turbine exhibits a more significant wake deflection relative to other downstream turbines. However, as the allowable power reduction increases, the positive effect of the yaw of the rear turbines on extending the fatigue life of the wind farm gradually becomes apparent, and larger yaw angles begin to appear. Figure 8 further quantifies and compares the actual power reduction and the resulting increase in fatigue life under different allowable power reduction coefficients. As shown in the figure, the change in actual power reduction always remains consistent with the allowable power reduction coefficient, while the resulting life extension factor shows a trend of first rapidly increasing and then gradually slowing down with the increase in allowable power reduction. For this embodiment... 0 = 98% is the dividing point, meaning that a 2% power reduction can achieve an average lifespan extension of nearly 8%, representing the maximum lifespan increase achievable per unit power reduction. This embodiment fully verifies that the multi-objective yaw control method for wind farms that considers the comprehensive benefits of power load can achieve a significant increase in the average lifespan of the wind farm with a relatively small power reduction, thus achieving the optimal comprehensive power load benefits of wind farm yaw control.

[0117] Of course, the above description is not intended to limit the present invention, and the present invention is not limited to the examples given above. Any changes, modifications, additions or substitutions made by those skilled in the art within the scope of the present invention should also fall within the protection scope of the present invention.

Claims

1. A multi-objective yaw control method for wind farms considering the comprehensive benefits of power load, characterized in that, The process includes the following steps: Step 1: Obtain the layout information of the wind farm, basic parameters of the wind turbines, and measured wind speed and direction data. Basic wind turbine parameters include: rotor diameter D, hub height H, power-wind speed curve, thrust coefficient curve, and the SN curve of the turbine transition section pipe node. Step 2: Establish a local three-dimensional coordinate system with the current wind flow direction as the positive x-axis, and determine the local three-dimensional coordinate data of each wind turbine in the wind farm based on the layout information obtained in Step 1. Step 3: Based on the measured wind speed and direction data obtained in Step 1 and the local three-dimensional coordinate data of the wind turbines obtained in Step 2, use the yaw analytical wake model to model the wake of the entire wind farm considering yaw conditions, and establish a whole-farm wake prediction model. Step 4: Establish the relationship between the wind farm yaw control strategy and the total power generation. Based on the wake prediction model established in step 3, the inflow wind speed of each wind turbine in the wind farm is determined. Combined with the power-wind speed curve, the power of each wind turbine is calculated, and the total power of the wind farm is obtained by summing them. Step 5: Establish the relationship between the wind farm yaw control strategy and the mean fatigue life. Based on the wake prediction model established in step 3, the inflow wind velocity and turbulence intensity of each wind turbine in the wind farm are determined. Combined with the SN curve of the pipe node in the transition section of the wind turbine, the fatigue life of each wind turbine is calculated, and the average value is taken to obtain the average life of the wind farm; Step 6: Based on the relationship between the wind farm yaw control strategy and the total power generation established in step 4. Step 7: Combining a Bayesian optimization learning network, the first stage of yaw control optimization considering a single objective function of power generation is performed, and the optimal yaw combination and corresponding total power are output as the baseline yaw strategy and baseline power. Simultaneously, using the relationship between the wind farm yaw control strategy and mean fatigue life established in step 5, the wind farm fatigue life under the baseline yaw strategy is calculated and used as the baseline mean life. The relationship between wind farm yaw control strategy and mean fatigue life Combining a Bayesian optimization learning network, with the lifetime extension factor relative to the baseline average lifetime as the optimization objective, and based on the actual power generation demand of the wind farm, the allowable reduction coefficient of the relative baseline power is set as a constraint to carry out the second stage of multi-objective yaw control optimization considering the comprehensive benefits of power load; Step 8: Output the final optimal yaw control strategy and its corresponding total power and average fatigue lifetime, and visualize the wake distribution of the entire field, thus completing the multi-objective yaw control of the wind farm considering the comprehensive benefits of power load.

2. The multi-objective yaw control method for wind farms considering comprehensive power load benefits according to claim 1, characterized in that, In step 1, the layout information of the wind farm is the global three-dimensional coordinate data of each wind turbine in a global three-dimensional coordinate system with due east as the positive x-axis direction.

3. The multi-objective yaw control method for wind farms considering the comprehensive benefits of power load according to claim 2, characterized in that, In step 2, determining the local three-dimensional coordinate data of each wind turbine in the wind farm based on the wind farm layout information obtained in step 1 specifically involves: converting the global three-dimensional coordinate data of each wind turbine obtained in step 1 into local three-dimensional coordinate data of each wind turbine in a local three-dimensional coordinate system; the conversion between the global three-dimensional coordinate data and the local three-dimensional coordinate data of the wind turbine satisfies the following relationship: Where (x, y) represents the global three-dimensional coordinate data of the wind turbine, and (x1, y1) represents the local three-dimensional coordinate data of the wind turbine. The angle between the local coordinate system and the global coordinate system in the counterclockwise direction.

4. The multi-objective yaw control method for wind farms considering the comprehensive benefits of power load according to claim 1, characterized in that, The process of establishing the overall wake prediction model in step 3 is as follows: Step 3.1: Number all wind turbines in ascending order of their horizontal coordinates (1, …, N), where N represents the number of wind turbines; Step 3.2: Based on the turbine numbers, determine the inflow wind speed and turbulence intensity of each wind turbine in the wind farm. The inflow wind speed and turbulence intensity of the first wind turbine are measured wind speed data. For the i-th downstream wind turbine, its inflow conditions are obtained by superimposing the wakes of all upstream wind turbines. The inflow wind speed and turbulence intensity are superimposed using the sum of squares and linear superposition methods, respectively; Step 3.3: Based on the inflow wind speed and turbulence intensity of each wind turbine, calculate the inflow wind speed and turbulence intensity at any position in the downstream wake region using a yaw analytical wake model that fully considers the effects of atmospheric turbulence intensity, thrust coefficient, and yaw angle. The velocity deficit and additional turbulence exhibit Gaussian and double Gaussian distributions radially, respectively, as shown in the following equation: ； ;in, For the inflow wind speed, For turbulence intensity, and For upstream undisturbed wind speed and turbulence intensity, The thrust coefficient to account for yaw effect, Where is the inflow turbulence intensity to the fan, D is the fan rotor diameter, and s is the wake diffusion distance. To account for the radial distance of the yaw wake centerline offset, K is the standard deviation of the Gaussian function, and k1 and k2 are the wake diffusion coefficients.

5. A multi-objective yaw control method for wind farms considering comprehensive power load benefits according to claim 1, characterized in that, The fatigue life calculation for each fan in step 5 specifically includes: Step 5.1: Calculate the peak stress based on the average wind speed and turbulence intensity of the incoming flow to the fan, combined with the fan thrust coefficient curve. and mean stress ： ;in, and These represent the average and peak wind speeds of the incoming airflow to the fan, respectively. Where is the pile radius, Let be the moment of inertia of the pile section, h be the height of the wind turbine hub, and A be the swept area of ​​the wind turbine. Given the air density; Step 5.2: Calculate the alternating stress amplitude based on the peak stress and average stress. To account for the influence of mean stress on the structural fatigue life, the Goodman curve was used to correct it, thus obtaining the equivalent alternating stress amplitude. ： ； ;in, For ultimate strength, For yield limit, To correct the mean stress; Step 5.3: Based on the SN curve of the pipe node in the transition section of the wind turbine, according to the equivalent alternating stress amplitude Calculate the corresponding fatigue life, i.e., the number of failure cycles N: ； Where t is the thickness of the area where fatigue cracks may grow. For reference thickness, the value is 32 mm for tubular nodes, and k is the thickness coefficient.

6. A multi-objective yaw control method for wind farms considering comprehensive power load benefits according to claim 1, characterized in that, In step 6, Bayesian machine learning is used to optimize yaw control in the first stage, considering a single objective function of power generation. Specifically, this includes: Step 6.1: Randomly generating a set of yaw control strategies, and using the relationship between the wind farm yaw control strategy and total power generation established in step 4. Step 6.2: Obtain the corresponding total power and generate an initial training set consisting of the yaw control strategy and the corresponding total power; Step 6.3: Use a Gaussian process to establish an approximate probability distribution between the yaw control strategy and the total power in the training set, and obtain the expectation and variance of the power prediction probability distribution; Step 6.4: Based on the obtained expectation and variance, establish a data acquisition function, search for possible optimal yaw combinations, and use the relationship between the wind farm yaw control strategy and the total power generation established in Step 4 to obtain the corresponding total power, forming new data points and incorporating them into the training set; Step 6.5: Use the training set obtained in Step 6.3 to repeatedly generate the power prediction probability distribution and search for possible optimal yaw control strategies until the optimization convergence condition is met.

7. A multi-objective yaw control method for wind farms considering comprehensive power load benefits according to claim 1, characterized in that, In step 7, the second stage of multi-objective yaw control optimization considering the comprehensive benefits of power load incorporates the power constraint considering the allowable reduction coefficient into the lifetime extension factor of the objective function in the form of a penalty function. Based on the relationship between the wind farm yaw control strategy and the total power generation established in step 4 and the relationship between the wind farm yaw control strategy and the average fatigue life established in step 5, a new relationship between the objective value and the yaw control strategy is established: Among them, L baseline and P baseline These are the baseline average lifetime and baseline power obtained in step 6, respectively. 0 represents the allowable power reduction factor. This is a penalty factor.

Images