A wind farm layout design method considering wind turbine fatigue control

Abstract

The present invention discloses a wind farm layout design method considering wind turbine fatigue control, which constructs a fatigue load database according to the load influencing factors of the wind turbine; based on the pre-constructed wake model and fatigue load database, the layout of the wind farm is optimized by using a machine learning method and a random search optimization algorithm; wherein the optimization design includes the fatigue load of the wind turbine and the annual power generation of the wind farm. The factors affecting the load size are taken into consideration, and the fatigue life of the wind turbine blades, towers, and shafts are taken into consideration to prevent the phenomenon that although the annual power generation of the wind farm increases after optimization, the fatigue damage of the wind turbines in the wind farm is far greater than the initial fatigue damage.

Description

Technical Field

[0001] The present application belongs to the technical field of wind farm power generation, and in particular relates to a wind farm layout design method considering wind turbine fatigue control. Background Art

[0002] Energy is closely related to social development. With the continuous exploitation of fossil energy such as coal, oil, and natural gas, the depletion of fossil energy is inevitable. Offshore wind power, as a new type of renewable energy, has become one of the main solutions to the shortage of fossil energy with its green, low-carbon and non-occupying land resources. Although the output efficiency of wind power is stable and the installed capacity is developing rapidly. At present, the layout optimization of wind farms is based on code programming and calculation considering the engineering wake model. Traditional programming calculations of wind farms only consider wind speed and direction, but rarely consider other issues, which makes the optimized layout often inconsistent with the actual situation. Therefore, considering the impact of wind loads on the service life of the structure under complex wake conditions in the optimization of wind turbine layout will make the layout optimization more realistic. Summary of the invention

[0003] The purpose of the present invention is to overcome the deficiencies in the prior art and provide a wind farm layout design method that takes into account wind turbine fatigue control, taking into account factors that affect the load size, and taking into account the fatigue life of wind turbine blades, towers, and shafts, to prevent the phenomenon that although the annual power generation of the wind farm increases after optimization, the fatigue damage of the wind turbines in the wind farm is far greater than the initial fatigue damage.

[0004] To achieve the above object, the present invention is implemented by adopting the following technical solutions:

[0005] In a first aspect, the present invention provides a wind farm layout design method considering wind turbine fatigue control, comprising:

[0006] Construct a fatigue load database based on the load influencing parameters of the wind turbine;

[0007] Based on the pre-built wake model and fatigue load database, the layout of the wind farm is optimized by using a random search optimization algorithm; wherein the optimization design includes the fatigue load of the wind turbine and the annual power generation of the wind farm.

[0008] Furthermore, the wake model is an improved Gaussian wake model, and the calculation formula of the wake model is:

[0009]

[0010] Where U hub is the inflow velocity at hub height, α is the wind shear exponent, F is the maximum wake velocity loss ratio, r is the distance from any point downstream to the wake center, Uw is the wake velocity, r w is the wake width, z hub represents the hub height, and z represents the vertical distance;

[0011] Among them, the parameter σ 2 g , A g and B g , the calculation formula is as follows:

[0012]

[0013] Among them, for the parameter r w The calculation formula is as follows:

[0014]

[0015] Where a = 1.216, b = 0.2071-0.144 / F, c = 0.2883(2 / F-1)-C T r d 2 / r 2 w F 2 , k = 0.4, x is the distance from the wind turbine in the downstream direction, D is the diameter of the wind turbine, r d is the wind turbine radius;

[0016] The calculation formula for parameter F is as follows:

[0017]

[0018] In the formula, Ti0 is the ambient turbulence intensity, C T is the thrust coefficient.

[0019] Furthermore, the step of constructing a fatigue load database according to the load influencing parameters of the wind turbine includes:

[0020] Analyze the factors affecting wind turbine load; calculate the wind turbine fatigue load database; process the wind turbine fatigue load database.

[0021] Furthermore, the load influencing factors of the wind turbine include: the instantaneous thrust of the wind turbine, the turbulence intensity and the wind speed;

[0022] The instantaneous thrust of the wind turbine is related to the instantaneous wind speed faced by the wind rotor, and the calculation formula is:

[0023]

[0024] Where A is the area of ​​the wind wheel, C is Tis the thrust coefficient, ρ is the air density, u0 is the instantaneous wind speed, and the instantaneous thrust of the wind turbine is proportional to the square of the instantaneous wind speed u0;

[0025] Wherein, the calculation formula of the turbulence intensity is:

[0026]

[0027] In the formula, u', v', and w' represent the fluctuating wind speed components in three orthogonal directions;

[0028] The wind speed is related to the vertical height, and the calculation formula is:

[0029]

[0030] Where U1 corresponds to the average wind speed at height z1, and α is the wind shear index.

[0031] Furthermore, the calculation of the wind turbine fatigue load database includes:

[0032] Use MATLAB, TurbSim and OpenFAST to generate wind field data required for simulating wind turbine operation, import the generated time series wind field data binary file into OpenFAST for simulation, and obtain the wind turbine operation data, wherein the wind turbine operation data includes power and torque;

[0033] The database selection includes the following loads: lateral bending moment, fore-aft bending moment and yaw moment at the tower base; rolling moment, tilting moment and yaw moment at the tower top; main axis bending moment; swing moment, flapping moment and torsional moment at the blade root.

[0034] Furthermore, the calculation process of the wind turbine fatigue load database specifically includes:

[0035] Calculate the SN curve, where the SN curve is the basis for structural strength analysis, representing the relationship between the cyclic stress S of a specific material and the number of failure cycles N. The calculation formula of the SN curve is:

[0036]

[0037] Where m and K are material parameters;

[0038] Calculate fatigue damage using the following formula:

[0039]

[0040] In the formula, S i is the load amplitude, n i For S iThe corresponding loop count;

[0041] Cycle N times in time T T Sub-constant load, the fatigue damage calculation formula is transformed into:

[0042]

[0043] Where DEL is the equivalent fatigue load, expressed as:

[0044]

[0045] Where N T Indicates N cycles within time T T Sub-constant load.

[0046] Further, the processing of the wind turbine fatigue load database includes: using a BP neural network to process the wind turbine fatigue load database;

[0047] The specific processing process is: divide the fatigue load data into training set and test set, and use the training set and test set to train the BP neural network; the input of the BP neural network is the average wind speed, turbulence intensity, wind shear index, inflow direction distance, and distance between axes, and the output is equivalent fatigue load data.

[0048] Furthermore, the layout of the wind farm is optimized by using a random search optimization algorithm based on a pre-built wake model and fatigue load database, including:

[0049] Calculate the power generation of the wind farm, represented by Weibull distribution:

[0050]

[0051] Among them, c is the shape coefficient, A is the proportionality coefficient, and v is the inflow wind speed;

[0052] The square sum model is used to calculate the wake superposition, and the calculation formula is:

[0053]

[0054] Where U0 is the average wind speed of the upstream wind turbine, U ji is the wake velocity from the jth wind turbine to the ith wind turbine, and n is the number of superimposed wakes;

[0055] The formula for calculating annual power generation is:

[0056]

[0057] Where AEP represents annual power generation, F(v k ,θw ) is the wind speed v k , wind direction is θ w The probability of (X, Y) is the wind farm layout, P tot () represents the actual power generation of the wind farm, N ws is the number of input wind speeds, N wd is the number of input wind speeds;

[0058] Incorporate the equivalent fatigue load of the wind turbine into the constraints:

[0059] (x i ,y i )∈S WT_area

[0060] 5D≤Dist Nt

[0061] S eq,DEL ≤Tol·S eq,DEL_ref

[0062] Among them, (x i ,y i ) is the specific location of the wind turbine, S WT_area is the boundary of the wind farm, Dist Nt is the minimum distance between any wind turbines, S eq,DEL In order to optimize the equivalent fatigue load of wind turbine blades, towers and shafts in the wind farm, S eq,DEL_ref is the equivalent fatigue load of wind turbine blades, towers and shafts in the initial layout of the wind farm. Tol is the tolerance coefficient or fitness of fatigue constraints, which is used to constrain the load size of wind turbines after layout optimization.

[0063] In a second aspect, the present invention discloses a computer device, including a processor and a storage medium;

[0064] The storage medium is used to store instructions;

[0065] The processor is used to operate according to the instruction to implement the steps of the wind farm layout design method considering wind turbine fatigue control as described in the first aspect.

[0066] In a third aspect, the present invention discloses a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the method for designing a wind farm layout taking into account wind turbine fatigue control as described in the first aspect is implemented.

[0067] Compared with the prior art, the present invention has the following beneficial effects:

[0068] The wind farm layout optimization method considering fatigue constraints of the present invention is an innovation of the traditional analytical method for solving wind farm layout optimization. The present invention takes into account the factors that affect the load size, and takes into account the fatigue life of wind turbine blades, towers, and shafts, to prevent the phenomenon that although the annual power generation of the wind farm increases after optimization, the fatigue damage of the wind turbines in the wind farm is far greater than the initial fatigue damage. This will make the wind farm optimization more in line with the actual needs of the project. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 It is a schematic flow chart of a method for optimizing the layout of a wind farm considering fatigue constraints according to an embodiment of the present invention;

[0070] Figure 2 The figure shows a schematic diagram of wind turbine flow generated by TurbSim according to an embodiment of the present invention;

[0071] Figure 3 It is a schematic diagram showing the relationship between the equivalent fatigue load, wind speed and turbulence intensity under normal incoming flow according to an embodiment of the present invention;

[0072] Figure 4 It is a schematic diagram showing the relationship between different fatigue loads and the distance x between upstream and downstream wind turbines when the turbulence intensity is 10% according to an embodiment of the present invention;

[0073] Figure 5 FIG. 1 is a schematic diagram of a partial training result of a BP neural network according to an embodiment of the present invention;

[0074] Figure 6 It is a schematic diagram of the process of optimizing a wind farm using a random search optimization algorithm according to an embodiment of the present invention;

[0075] Figure 7 Shown is a schematic diagram of an initial layout of a wind farm used for research according to an embodiment of the present invention;

[0076] Figure 8 It is a schematic diagram showing the comparison of the layout of a wind farm with all-wind direction flow before and after optimization according to an embodiment of the present invention;

[0077] Fig. 9 The figure shows a schematic diagram of the AEP optimization result of a wind farm with full wind direction flow according to an embodiment of the present invention;

[0078] Fig.10 It is a schematic diagram of fatigue load MyST after the layout of the wind farm with full wind flow is optimized according to one embodiment of the present invention;

[0079] Fig.11 Shown is a schematic diagram of the technical architecture of wind farm layout optimization software considering wind turbine fatigue control according to an embodiment of the present invention. DETAILED DESCRIPTION

[0080] The technical concept of the present invention is: construct a fatigue load database according to the load influencing parameters of the wind turbine; based on the pre-constructed wake model and fatigue load database, use a random search optimization algorithm to optimize the layout of the wind farm; wherein the optimization design includes the fatigue load of the wind turbine and the annual power generation of the wind farm. The factors affecting the load size are taken into consideration, and the fatigue life of the wind turbine blades, towers, and shafts are taken into consideration to prevent the phenomenon that although the annual power generation of the wind farm increases after optimization, the fatigue damage of the wind turbines in the wind farm is far greater than the initial fatigue damage.

[0081] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. The following description of at least one exemplary embodiment is actually only illustrative and is by no means intended to limit the present application and its application or use.

[0082] Example 1

[0083] Traditional wind farm layout optimization methods often only calculate based on simple parameters such as wind speed and wind direction, ignoring many other complex factors. This approach results in the optimized wind farm layout not meeting actual operational requirements. Therefore, this embodiment provides a wind farm layout design method that takes into account wind turbine fatigue control, such as Figure 1 As shown in the figure, this method takes into account the impact of wind loads on the service life of wind turbine structures in complex situations. This consideration will make the layout optimization of wind farms closer to actual operating conditions, thereby improving the safety and economy of wind farms. This example includes the following steps:

[0084] Construct a fatigue load database based on the load influencing parameters of the wind turbine;

[0085] Based on the pre-built wake model and fatigue load database, the layout of the wind farm is optimized by using a random search optimization algorithm; wherein the optimization design includes the fatigue load of the wind turbine and the annual power generation of the wind farm.

[0086] In this embodiment, an improved Gaussian wake model is selected as the wake model for wind farm layout optimization, and the calculation formula is as follows:

[0087]

[0088] Among them U hub is the inflow velocity at hub height, α is the wind shear exponent, F is the maximum wake velocity loss ratio, r is the distance from any point downstream to the wake center, U w is the wake velocity, r w is the wake width, z hubrepresents the hub height, and z represents the vertical distance.

[0089] For the parameter σ 2 g , A g and B g , the calculation formula is as follows:

[0090]

[0091] For the parameter r w The calculation formula is as follows:

[0092]

[0093] Where a = 1.216, b = 0.2071-0.144 / F, c = 0.2883(2 / F-1)-C T r d 2 / r 2 w F 2 , r d is the radius of the wind turbine, k = 0.4, x is the distance from the wind turbine in the downstream direction, and D is the diameter of the wind turbine.

[0094] The calculation formula for parameter F is as follows:

[0095]

[0096] Among them, Ti0 is the ambient turbulence intensity, C T is the thrust coefficient.

[0097] The process of analyzing the influencing factors related to the load database is as follows:

[0098] In the field of wind farm layout optimization, the relevant simulation calculation time interval is generally 10 minutes, and the height of the wind turbine hub is used as the average wind speed reference height. The instantaneous thrust of the wind turbine is related to the instantaneous wind speed faced by the wind rotor. The calculation formula is as follows:

[0099]

[0100] Where A is the area of ​​the wind wheel, C T is the thrust coefficient, ρ is the air density, and u0 is the instantaneous wind speed. The instantaneous thrust of the wind turbine is proportional to the square of the instantaneous wind speed u0. The load of the wind turbine is related to the thrust. Therefore, the average wind speed has a great influence on the load of the wind turbine. The average wind speed is a crucial influencing factor in the load calculation.

[0101] Turbulence intensity describes the relative intensity of wind speed fluctuations within a certain period of time. The calculation formula is as follows:

[0102]

[0103] Where u', v', and w' represent the fluctuating wind speed components in three orthogonal directions. The time series of incoming flow with high turbulence intensity usually leads to a large degree of fluctuation in wind turbine load. Therefore, the turbulence intensity has a strong influence on the amplitude of wind turbine load fluctuation.

[0104] The magnitude of wind speed is related to vertical height. This phenomenon is called wind shear and is generally described by an exponential function. The calculation formula is as follows:

[0105]

[0106] Among them, U1 corresponds to the average wind speed at height z1, and α is the wind shear index. Since the height span of each point on the wind rotor plane is large, the wind speed varies greatly up and down the wind rotor, the thrust in the upper part is large, and the thrust in the lower part is small. Therefore, the influence of wind shear needs to be considered when calculating the wind turbine load.

[0107] Since the calculated load of wind turbines in a wind farm needs to take into account the wake effect between wind turbines in the entire wind farm, it is necessary to use the Gaussian distribution wake model and the wake meandering theory to calculate the wind turbine load in the wake area. The specific calculation parameters are the distance x between the downstream wind turbine and the upstream wind turbine in the inflow direction, and the distance y between the axis of the downstream wind turbine and the axis of the upstream wind turbine.

[0108] The process of calculating the wind turbine load database is as follows:

[0109] MATLAB, TurbSim and OpenFAST are used to generate the wind field data required to simulate the operation of the wind turbine. The generated binary files of the time series wind field data (u, v, w) are then imported into OpenFAST for simulation, and finally the operation data of the wind turbine, including power, torque, etc., are obtained.

[0110] The database selection includes the following 11 loads: lateral bending moment, fore-aft bending moment and yaw moment at the base of the tower (MxTB, MyTB and MzTB); rolling moment, tilting moment and yaw moment at the top of the tower (MxTT, MyTT and MzTT); main axis bending moment of rotation (MyST and MzST); swing moment, flapping moment and torsional moment at the root of the blade (MxBR, MyBR and MzBR).

[0111] The process of calculating the wind turbine fatigue database is as follows:

[0112] The SN curve is an important basis for structural strength analysis. It represents the relationship between the cyclic stress S of a specific material and the number of failure cycles N. The calculation formula is as follows:

[0113]

[0114] Where m and K are material parameters.

[0115] In wind farm optimization applications, fatigue damage is usually expressed by equivalent fatigue loads, using processed fixed loads instead of random loads that fluctuate over time. The calculation formula is:

[0116]

[0117] The idea is to use a time series to calculate a load amplitude S through the rain flow counting method. i and the corresponding loop count n i The damage is calculated by combining the SN curve and linear fatigue cumulative damage theory.

[0118] Now assume that this section of damage is entirely caused by a cyclic load, that is, N cycles in time T T The sub-constant load is calculated as:

[0119]

[0120] In order to make the calculation more convenient and faster, the equivalent fatigue load DEL is often used, and the calculation formula is:

[0121]

[0122] In the simulation, since the simulation time is much shorter than the actual life, this paper takes a simulation time of 10 minutes (600S) to calculate the equivalent fatigue load of 1Hz, where the number of load cycles N in the 10-minute time series is T is 600.

[0123] The process of processing the wind turbine fatigue database is as follows:

[0124] Since the changes in DEL are difficult to describe and express in a single way, BP neural network is used for database processing. BP neural network is a multi-layer feedforward neural network trained using error back propagation. Its purpose is to learn specific rules through self-training so that when input values ​​are provided, the results are as close as possible to the expected output values.

[0125] The present invention performs model training and testing on each equivalent fatigue load in turn. Due to the large sample size of the database, most of the data are randomly selected for training, and the remaining 200 groups of data are used as test sets for testing. The inputs for prediction are the average wind speed U, turbulence intensity Ti, wind shear index α, inflow direction distance x, and axis-to-axis distance y. The predicted outputs are 11 equivalent fatigue loads.

[0126] The optimization process of wind farm layout considering fatigue load is as follows:

[0127] During the operation of a wind farm, it is necessary to know the inflow wind speed of each wind turbine in order to accurately estimate the power generation of the wind farm. Usually, after measuring a sufficient amount of wind data, the wind data is fitted to a probability distribution in a given direction, represented by the Weibull distribution:

[0128]

[0129] Where c and A are shape coefficient and proportional coefficient respectively, and v represents the inflow wind speed.

[0130] An improved Gaussian wake model in the wind farm introduced above is used as the wake model for wind farm layout optimization to deal with the non-negligible wake effect in the wind farm.

[0131] In an operating wind farm, downstream wind turbines will inevitably be affected by multiple upstream wind turbines, so the problem of wake mixing and superposition must be considered. The present invention adopts a widely used and highly recognized square sum model to calculate the wake superposition, and the calculation formula is:

[0132]

[0133] Where U0 is the average wind speed of the upstream wind turbine, U ji is the wake velocity from the jth wind turbine to the ith wind turbine, and n is the number of superimposed wakes.

[0134] The turbulence intensity of normal inflow is calculated according to the normal turbulence model, and the calculation formula is:

[0135]

[0136] Among them I ref is the turbulence intensity when the hub height is 15m / s, which is determined according to the safety level of the unit and is taken as 0.14 in this paper.

[0137] The turbulence intensity in the wake zone is also affected as is the wind speed, and is calculated as:

[0138]

[0139] Where I0 is the ambient turbulence intensity, I + is the wake additional turbulence intensity, calculated as:

[0140]

[0141] Among them I + , m is the maximum wake additional intensity, and the calculation formula is:

[0142]

[0143] The value of k2 is taken as 0.5, x is the distance from the wind turbine in the downstream direction, D is the diameter of the wind turbine, and C T is the thrust coefficient.

[0144] The superposition of turbulence intensity in the wake mixing zone is similar to the superposition of wind speed, and the calculation formula is:

[0145]

[0146] Where ΔI ji It represents the impact of the j-th upstream wind turbine on the i-th downstream wind turbine.

[0147] For the same type of wind turbine, the manufacturer usually provides a power curve and a thrust coefficient curve. In different situations, the power of the wind turbine can be directly obtained from the wind speed in front of the wind rotor. After the wind farm layout (X, Y), wind speed and wind direction are known, the calculation formula for annual power generation AEP is:

[0148]

[0149] Among them, F(v k ,θ w ) is the wind speed v k , wind direction is θ w The probability of tot () represents the actual power generation of the wind farm, N ws is the number of input wind speeds, N wd To input the number of wind speeds, assume that the annual operating time of the wind farm is 8760 hours.

[0150] The optimization of wind farms is subject to various restrictions. Usually, two constraints are considered in layout optimization. One is the wind farm boundary, and the other is the minimum distance between wind turbines. The layout optimization method of the present invention incorporates the equivalent fatigue load of wind turbines into the constraints to form three sets of constraints:

[0151] (x i ,y i )∈S WT_area

[0152] 5D≤Dist Nt

[0153] S eq,DEL ≤Tol·S eq,DEL_ref

[0154] Among them, (x i ,y i ) is the specific location of the wind turbine, S WT_area is the boundary of the wind farm, Dist Nt is the minimum distance between any wind turbines, Seq,DEL In order to optimize the equivalent fatigue load of wind turbine blades, towers and shafts in the wind farm, S eq,DEL_ref is the equivalent fatigue load of wind turbine blades, towers and shafts in the initial layout of the wind farm. Tol is the tolerance coefficient or fitness of fatigue constraints, which is used to constrain the load size of wind turbines after layout optimization.

[0155] In the intelligent optimization algorithm, this embodiment adopts a random search optimization algorithm, which is a wind farm layout optimization algorithm based on the continuity formula. The principle is that at the beginning of the optimization, the algorithm will record the initial position S0, annual power generation AEP0 and equivalent fatigue load DELs0 of the wind farm and set the failure of annual power generation to 'False' and success to 'True'. After that, the random search optimization algorithm will randomly select any one of the wind farms and move it by a random step. After each step, the position after the move will be detected to observe whether it still stays within the set wind turbine layout range. If the detection result is no, then the AEP of the wind farm in this step is recorded as zero, and the first step is re-entered to select the wind turbine that was moved last time to move a random step again. When it is determined to be within the selected range after this movement, the AEP of the wind farm at this position will be calculated. If the AEP is improved, the step will be retained and recorded, and the wind turbine will continue to be selected to move a random distance until the AEP is not improved. At this time, other wind turbines are selected to repeat the above operation. Finally, we can set the number of iterations we need, such as 30,000 or 40,000 iterations, until the optimized AEP becomes stable.

[0156] The following is a specific example to illustrate the wind farm layout design method considering wind turbine fatigue control in this embodiment, which specifically includes the following steps:

[0157] Step 1: Construct a simulation scheme based on the parameters affecting the load value determined in the technical solution, and generate the inflow wind required by the wind turbine in Turbsim. It is also necessary to use the improved Gaussian wake model and wake meander model mentioned above to calculate the wind turbine load in the wake area.

[0158] Among them, the average wind speed at the hub height is U hub From 3m / s to 12m / s in 1m / s intervals, from 13m / s to 25m / s in 2m / s intervals.

[0159] The turbulence intensity Ti at the hub height is in the range of 10% to 35% at intervals of 2.5% according to the unit design range of the normal turbulence model.

[0160] The distance x between the downstream wind turbine and the upstream wind turbine in the inflow direction ranges from 3D to 10D at intervals of 1D, and from 12D to 20D at intervals of 2D. (D is the diameter of the wind turbine)

[0161] Among them, the distance y between the axis of the downstream wind turbine and the axis of the upstream wind turbine are 0D, 0.25D, 0.5D, 1D, 1.5D and 2D respectively.

[0162] The wind shear index and wind turbine yaw angle are set to 0.1 and 0° respectively. In addition, the turbulence model selected for the inflow wind generated in TurbSim is the Kaimal model.

[0163] The number of cases selected to construct the fatigue load database includes 187 (17×11) upstream wind turbine cases that are not affected by wake and 14586 (17×11×13×6×1×1) downstream wind turbine cases that include wake effects.

[0164] Figure 2 The schematic diagram of the wind turbine inflow generated by TurbSim provided by the present invention. According to the normal inflow and the velocity loss and turbulence intensity in the wake area, the inflow wind field data required for the operation of the wind turbine at any position in the wind farm can be constructed, and it is imported into OpenFAST as a binary file for simulation. The simulation time is set to 630s each time, and the results are analyzed after 600s, and finally the operation database of all case wind turbines is obtained, which includes the 11 loads mentioned above.

[0165] Step 2: After a long period of simulation, a large database containing various operating environments, such as different average wind speeds, turbulence intensity and wake effects, is obtained. Then the DEL calculation method introduced above is used to calculate the DEL of 11 loads. Part of the data of the fatigue load database calculated by MATLAB, TurbSim and OpenFAST in this invention is shown in Table 1.

[0166] Case 1 ··· 187 188 14773 Average wind speed 3 ··· 25 3 25 Turbulence intensity 10％ ··· 35％ 10％ 35％ x-direction position - ··· - 3D 20D Y position - ··· - 0D 2D DELMxTB 6.3021E+02 ··· 3.7503E+04 4.5604E+02 3.7513E+04 DELMxTB 4.2477E+03 ··· 5.7089E+04 2.2692E+03 5.7016E+04 ··· ··· ··· ··· ··· ··· DELMxTB 3.5307E+03 ··· 1.6810E+04 3.4092E+03 1.6776E+04 DELMxTB 3.5393E+03 ··· 1.6994E+04 3.5417E+03 1.6924E+04

[0167] Table 1

[0168] Figure 3 The present invention provides a partial correlation diagram between equivalent fatigue load and average wind speed and turbulence intensity, from which it can be seen that with the increase of wind speed or turbulence intensity, the equivalent fatigue load of the wind turbine will be affected and increase.

[0169] Figure 4 The relationship diagram between different fatigue loads and the upstream and downstream wind turbine spacing x when the turbulence intensity is 10% provided by the present invention shows that most equivalent fatigue loads increase slowly with the increase of downstream distance, but some equivalent fatigue loads decrease or fluctuate with the increase of downstream distance. Since the equivalent fatigue changes are different and difficult to express with a single method, machine learning methods are needed for database processing.

[0170] Step 3: After constructing the BP neural network parameters, such as the number of network layers, the number of neurons in each layer, the activation function and the training function, the model training and testing are performed on the above 11 equivalent fatigue load databases. Most of the data are randomly selected for training, and 200 groups of data are reserved as test sets for testing.

[0171] Figure 5 This is a partial training result diagram of the BP neural network provided by the present invention. It can be seen that the predicted values ​​of the four equivalent fatigue loads listed after BP neural network training are almost perfectly consistent with the simulation results, and the relative errors are all within ±5%.

[0172] Step 4: Program the random search algorithm mentioned above in MATLAB, where Figure 6 The flowchart of the random search optimization algorithm provided by the present invention for optimizing a wind farm is provided, and four layout optimization schemes without fatigue constraints and with fatigue values ​​lower than 1, 1.02 and 1.05 times of the initial layout are calculated by increasing the fitness Tol=None, 1.0, 1.02, 1.05.

[0173] Figure 7 The initial layout of the wind farm for case study provided by the present invention, wherein the wind turbine used is a NREL 5MW wind turbine. The AEP of the wind farm is calculated according to the improved Gaussian wake model calculation formula, turbulence model calculation formula and annual power generation AEP calculation formula mentioned above, and the layout is optimized according to the three major constraints (area constraint, spacing constraint and fatigue constraint) using a random search algorithm.

[0174] Figure 8 This is a comparison diagram before and after the optimization of the layout of the wind farm with all-wind direction inflow provided by the present invention. It can be seen that when the fatigue constraint limit becomes more and more stringent, the movement range of each wind turbine becomes smaller and smaller. Fig. 9 The AEP optimization result diagram of the wind farm with all wind directions provided by the present invention shows that the AEP increases of the four optimization results compared with the initial results are 7.48E+04MWh, 5.52E+04MWh, 4.18E+04MWh and 1.41E+04MWh respectively. The AEP obtained by ignoring the equivalent fatigue constraint optimization is the largest. As the fatigue constraint becomes more and more stringent, the AEP gradually decreases, but there is still an AEP increment that cannot be ignored. Therefore, it is feasible for the present invention to consider fatigue to extend the service life of the wind turbine.

[0175] Fig.10This is a comparison chart of fatigue load MyST after the layout optimization of the wind farm with full wind direction provided by the present invention. It can be clearly seen that MyST has more super-high fatigue loads in the layout optimization without fatigue constraints, while the equivalent fatigue load is greatly reduced after adding the fatigue constraint of Tol=1.05. When the restriction requirements are more stringent, when Tol=1.0, the equivalent fatigue load decreases more significantly.

[0176] Fig.11 The technical architecture diagram of the wind farm layout optimization software considering wind turbine fatigue control provided by the present invention. The design is carried out by adopting the B / S architecture, with the Django framework as the core, to build a connected backend (wind farm layout optimization model), front end (user interaction interface), and its own database (wind turbine fatigue load database). It is divided into an application server layer, a database server layer, and a computer browser layer. Business services such as wind farm optimization design algorithms are deployed on the application server, and the information database and business logic database are deployed on the database server, and the two communicate through the network. The computer browser layer is developed using the open source popular framework Bootstrap, and the interactive interface is friendly and easy to use, allowing users to perform related operations such as wind farm layout optimization considering wind turbine fatigue control.

[0177] Example 2

[0178] This embodiment discloses a computer device, including a processor and a storage medium;

[0179] The storage medium is used to store instructions;

[0180] The processor is used to operate according to the instruction to implement the steps of the wind farm layout design method considering wind turbine fatigue control as described in Example 1.

[0181] Example 3

[0182] This embodiment discloses a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for designing a wind farm layout taking into account wind turbine fatigue control as described in Embodiment 1 is implemented.

[0183] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0184] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0185] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0186] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0187] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A wind farm layout design method considering wind turbine fatigue control, characterized in that: include: A fatigue load database is constructed based on the load influencing parameters of the wind turbine, including: Analyze the factors affecting wind turbine load; calculate the wind turbine fatigue load database; process the wind turbine fatigue load database; The factors affecting the load of the wind turbine include: the instantaneous thrust of the wind turbine, the turbulence intensity and the wind speed; The wind turbine fatigue load calculation database includes: Use MATLAB, TurbSim and OpenFAST to generate the wind field data required for simulating the operation of wind turbines. Import the generated binary files of wind field data related to time series into OpenFAST for simulation to obtain the operation data of wind turbines. The processing of the wind turbine fatigue load database comprises: using a BP neural network to process the wind turbine fatigue load database; The specific processing process is as follows: the fatigue load data is divided into a training set and a test set, and the training set and the test set are used to train the BP neural network; the input of the BP neural network is the average wind speed, turbulence intensity, wind shear index, inflow direction distance, and distance between axes, and the output is equivalent fatigue load data; Based on the pre-built wake model and fatigue load database, the layout of the wind farm is optimized by using a random search optimization algorithm; wherein the optimization design includes the fatigue load of the wind turbine and the annual power generation of the wind farm; The wake model is an improved Gaussian wake model, and the calculation formula of the wake model is: Where U hub is the inflow velocity at hub height, α is the wind shear exponent, F is the maximum wake velocity loss ratio, r is the distance from any point downstream to the wake center, U w is the wake velocity, r w is the wake width, z hub represents the hub height, and z represents the vertical distance; Among them, the parameter σ 2 g , A g and B g , the calculation formula is as follows: Among them, for the parameter r w The calculation formula is as follows: Where a = 1.216, b = 0.2071-0.144 / F, c = 0.2883(2 / F-1)-C T r d 2 / r 2 w F 2 , k = 0.4, x is the distance from the wind turbine in the downstream direction, D is the diameter of the wind turbine, r d is the wind turbine radius; The calculation formula for parameter F is as follows: In the formula, Ti0 is the ambient turbulence intensity, C T is the thrust coefficient.

2. The wind farm layout design method considering wind turbine fatigue control according to claim 1 is characterized in that: The instantaneous thrust of the wind turbine is related to the instantaneous wind speed faced by the wind rotor, and the calculation formula is: Where A is the area of ​​the wind wheel, C is T is the thrust coefficient, ρ is the air density, u0 is the instantaneous wind speed, and the instantaneous thrust of the wind turbine is proportional to the square of the instantaneous wind speed u0; The calculation formula of the turbulence intensity is: In the formula, u', v', and w' represent the fluctuating wind speed components in three orthogonal directions; The wind speed is related to the vertical height, and the calculation formula is: Where U1 corresponds to the average wind speed at height z1, and α is the wind shear index.

3. The wind farm layout design method considering wind turbine fatigue control according to claim 1 is characterized in that: The operation data of the wind turbine includes power and torque; The database includes the following loads: lateral bending moment, fore-aft bending moment and yaw moment at the bottom of the tower; rolling moment, tilting moment and yaw moment at the top of the tower; rotating main axis bending moment; swing moment, flapping moment and torsional moment at the root of the blade.

4. The wind farm layout design method considering wind turbine fatigue control according to claim 3 is characterized in that: The calculation process of the wind turbine fatigue load database specifically includes: Calculate the SN curve, where the SN curve is the basis for structural strength analysis, representing the relationship between the cyclic stress S of a specific material and the number of failure cycles N. The calculation formula of the SN curve is: Where m and K are material parameters; Calculate fatigue damage using the following formula: In the formula, S i is the load amplitude, n i For S i The corresponding cycle count; Cycle N times in time T T Sub-constant load, the fatigue damage calculation formula is transformed into: Where DEL is the equivalent fatigue load, expressed as: Where N T Indicates N cycles within time T T Sub-constant load.

5. The wind farm layout design method considering wind turbine fatigue control according to claim 1, characterized in that: The method of optimizing the layout of the wind farm using a random search optimization algorithm based on a pre-built wake model and fatigue load database includes: Calculate the power generation of the wind farm, represented by Weibull distribution: Where c is the shape coefficient, A is the proportionality coefficient, and v is the inflow velocity; The square sum model is used to calculate the wake superposition, and the calculation formula is: Where U0 is the average wind speed of the upstream wind turbine, U ji is the wake velocity from the jth wind turbine to the ith wind turbine, and n is the number of superimposed wakes; The annual power generation calculation formula is: In the formula, AEP represents annual power generation, F(v k ,θ w ) is the wind speed v k , wind direction is θ w The probability of (X, Y) is the wind farm layout, P tot () represents the actual power generation of the wind farm, N ws is the number of input wind speeds, N wd is the number of input wind speeds; Incorporate the equivalent fatigue load of the wind turbine into the constraints: (x i ,and i )∈S WT_area 5D≤Dist Nt S eq,DEL ≤Tol·S eq,DEL_ref Among them, (x i ,y i ) is the specific location of the wind turbine, S WT_area is the boundary of the wind farm, Dist Nt is the minimum distance between any wind turbines, S eq,DEL In order to optimize the equivalent fatigue load of wind turbine blades, towers and shafts in the wind farm, S eq,DEL_ref is the equivalent fatigue load of wind turbine blades, towers and shafts in the initial layout of the wind farm. Tol is the tolerance coefficient or fitness of fatigue constraints, which is used to constrain the load size of wind turbines after layout optimization.

6. A computer device, characterized in that: including processor and storage medium; The storage medium is used to store instructions; The processor is used to operate according to the instructions to implement the steps of the wind farm layout design method considering wind turbine fatigue control as described in any one of claims 1 to 5.

7. A computer storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed by a processor, the wind farm layout design method considering wind turbine fatigue control as described in any one of claims 1 to 5 is implemented.

Images