Fatigue load suppression method, device, equipment and medium in wake wind farm

Abstract

The present application discloses a method, device, equipment and medium for suppressing fatigue load in a wake wind farm, the method comprising: using the historical wind speed of the target unit in the wake wind farm to train a wind speed prediction model; using the wind speed prediction model to process the current wind speed of the target unit to obtain the predicted wind speed of the target unit; building a state space model of the wind turbine unit, and building a dynamic wake model in combination with the predicted wind speed and a preset wake sensitivity coefficient; combining the dynamic wake model with the state space model to establish a discretized equation; establishing a linearized expression for tower bending moment and shaft torque, and establishing an optimization objective function for tower bending moment and shaft torque based on the linearized expression and the discretized equation; solving the optimization objective function under preset constraints to obtain the optimal torque and the optimal pitch angle, and realizing fatigue load suppression based on the optimal torque and the optimal pitch angle. This scheme can suppress fatigue load in a wake wind farm more efficiently and accurately.

Description

Technical Field

[0001] The present application relates to the technical field of wind farms, and in particular to a fatigue load suppression method in a wake wind farm, and also to a fatigue load suppression device, electronic equipment and computer-readable storage medium in a wake wind farm. Background Art

[0002] Offshore wind farms are growing rapidly in number and size. To achieve better economic benefits, reducing maintenance costs is one of the most important factors. The fatigue loads on wind turbines have a significant impact on their maintenance costs. Reducing fatigue loads can produce less damage, which means less maintenance costs.

[0003] In fact, the fatigue load of wind turbines is jointly affected by wind speed and the active power command of the central controller of the wind farm. Due to the uncertainty and variability of wind energy, the variation characteristics of fatigue loads become more complicated. At present, the fatigue load of the unit can be reduced by predicting wind speed through radar and taking feedforward control, but it will increase the corresponding equipment cost. At the same time, large-scale wind farms have complex wake effects, and the existing numerical wake model has a long calculation time and cannot meet the needs of real-time load control.

[0004] Therefore, how to more efficiently and accurately suppress the fatigue load in the wake wind farm is an urgent problem to be solved by those skilled in the art. Summary of the invention

[0005] The purpose of the present application is to provide a method for suppressing fatigue loads in a wake wind farm, which can suppress the fatigue loads in the wake wind farm more efficiently and accurately; another purpose of the present application is to provide a fatigue load suppression device, electronic device and computer-readable storage medium in a wake wind farm, all of which have the above-mentioned beneficial effects.

[0006] In a first aspect, the present application provides a method for suppressing fatigue loads in a wake wind farm, comprising:

[0007] Obtaining historical wind speeds of target units in the wake wind farm, and using the historical wind speeds to train a wind speed prediction model;

[0008] Processing the current wind speed of the target unit using the wind speed prediction model to obtain the predicted wind speed of the target unit;

[0009] Building a state space model of the wind turbine, and building a dynamic wake model in combination with the predicted wind speed and the preset wake sensitivity coefficient;

[0010] Combining the dynamic wake model with the state space model to establish a discretized equation;

[0011] Establishing a linearized expression for tower bending moment and shaft torque, and establishing an optimization objective function for tower bending moment and shaft torque based on the linearized expression and the discretized equation;

[0012] The optimization objective function is solved under preset constraint conditions to obtain an optimal torque and an optimal pitch angle, and fatigue load suppression is achieved based on the optimal torque and the optimal pitch angle.

[0013] Optionally, the historical wind speed of the target unit in the wake wind farm is obtained, including:

[0014] At a preset time interval, collecting a first preset number of the historical wind speeds of the target unit in the wake wind farm;

[0015] Accordingly, the current wind speed of the target unit is processed by using the wind speed prediction model to obtain the predicted wind speed of the target unit, including:

[0016] The wind speed prediction model is used to process the current wind speed of the target unit to obtain the predicted wind speed at each preset time interval within the prediction time domain.

[0017] Optionally, a state space model of the wind turbine is constructed, and a dynamic wake model is constructed in combination with the predicted wind speed and a preset wake sensitivity coefficient, including:

[0018] Building a state space model of the wind turbine set, and determining a wind speed increment and a mechanical state increment of a wind turbine in the wake wind farm according to the state space model;

[0019] The dynamic wake model is constructed according to the wind speed increment of the wind turbine, the mechanical state increment of the wind turbine, the predicted wind speed, and the preset wake sensitivity coefficient.

[0020] Optionally, a state space model of the wind turbine is constructed, including:

[0021] Constructing a transmission system model, a pitch angle control system model, and a motor torque response system model of the wind turbine generator set according to various index parameters of the wind turbine generator set;

[0022] The state space model of the wind turbine generator set is constructed by using the transmission system model, the pitch angle control system model and the motor torque response system model.

[0023] Optionally, constructing the dynamic wake model according to the wind speed increment of the wind turbine, the mechanical state increment of the wind turbine, the predicted wind speed, and the preset wake sensitivity coefficient includes:

[0024] Determining a time delay for wake propagation based on the unit spacing of the wind turbine units;

[0025] A dynamic wake model based on wake time lag is constructed according to the wind speed increment of the wind turbine, the mechanical state increment of the wind turbine, the predicted wind speed, the preset wake sensitivity coefficient, and the time delay.

[0026] Optionally, the preset wake sensitivity coefficient includes:

[0027] The sensitivity coefficient of the wind turbine wake wind speed in the wind turbine group to the corresponding wind turbine wind speed;

[0028] The sensitivity coefficient of the wind turbine wake wind speed in the wind turbine generator set to the corresponding wind turbine speed;

[0029] The sensitivity coefficient of the wind turbine wake wind speed in the wind turbine generator set to the corresponding wind turbine pitch angle.

[0030] Optionally, the preset constraint condition includes:

[0031] The motor torque in the wind turbine generator set does not exceed the maximum motor torque;

[0032] The motor pitch angle in the wind turbine set does not exceed the maximum motor pitch angle;

[0033] The change in the motor torque in the wind turbine generator set satisfies a first preset range;

[0034] The variation of the motor pitch angle in the wind turbine generator set satisfies a second preset range.

[0035] In a second aspect, the present application also discloses a fatigue load suppression device in a wake wind farm, comprising:

[0036] A training module, used to obtain the historical wind speed of the target unit in the wake wind farm, and use the historical wind speed to train a wind speed prediction model;

[0037] A prediction module, used to process the current wind speed of the target unit using the wind speed prediction model to obtain the predicted wind speed of the target unit;

[0038] A first building module is used to build a state space model of the wind turbine and to construct a dynamic wake model in combination with the predicted wind speed and the preset wake sensitivity coefficient;

[0039] A second building module is used to combine the dynamic wake model with the state space model to establish a discretized equation;

[0040] A third establishing module is used to establish a linearized expression about the tower bending moment and the shaft torque, and to establish an optimization objective function about the tower bending moment and the shaft torque according to the linearized expression and the discretized equation;

[0041] A solution module is used to solve the optimization objective function under preset constraints to obtain an optimal torque and an optimal pitch angle, and to achieve fatigue load suppression according to the optimal torque and the optimal pitch angle.

[0042] In a third aspect, the present application further discloses an electronic device, comprising:

[0043] Memory for storing computer programs;

[0044] A processor is used to implement the steps of any one of the above-mentioned methods for suppressing fatigue loads in a wake wind farm when executing the computer program.

[0045] In a fourth aspect, the present application further discloses a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of any of the fatigue load suppression methods in a wake wind farm as described above are implemented.

[0046] The present application provides a fatigue load suppression method in a wake wind farm, comprising: obtaining a historical wind speed of a target unit in the wake wind farm, and using the historical wind speed to train a wind speed prediction model; using the wind speed prediction model to process the current wind speed of the target unit to obtain a predicted wind speed of the target unit; building a state space model of the wind turbine unit, and constructing a dynamic wake model in combination with the predicted wind speed and a preset wake sensitivity coefficient; combining the dynamic wake model with the state space model to establish a discretized equation; establishing a linearized expression for tower bending moment and shaft torque, and establishing an optimization objective function for tower bending moment and shaft torque based on the linearized expression and the discretized equation; solving the optimization objective function under preset constraints to obtain an optimal torque and an optimal pitch angle, and achieving fatigue load suppression based on the optimal torque and the optimal pitch angle.

[0047] By applying the technical solution provided in this application, the model is trained through the historical incoming wind speed in the wake wind farm to obtain the predicted value of the incoming wind speed in the future. The predicted wind speed can be used to estimate the future changes of fatigue load and take action in advance to achieve fatigue load suppression; and this solution proposes fatigue load suppression based on the dynamic wake model, which simplifies the traditional nonlinear, coupled analytical wake model into a linear, decoupled model, which is suitable for fast wake calculation; combined with the internal dynamic characteristics of the wind turbine, it can predict the wake wind speed loss under the change of the mechanical state (generator torque and pitch angle) of the wind turbine in the future, and then achieve fatigue load suppression on the tower and blades. It can be seen that this technical solution achieves more efficient and accurate fatigue load suppression in the wake wind farm.

[0048] The fatigue load suppression device, electronic device and computer-readable storage medium in the wake wind farm provided in the present application also have the above-mentioned technical effects, which will not be described in detail in the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the technical solutions in the prior art and the embodiments of the present application, the drawings required for use in the description of the prior art and the embodiments of the present application are briefly introduced below. Of course, the drawings described below in relation to the embodiments of the present application are only part of the embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without creative work, and the obtained other drawings also belong to the protection scope of the present application.

[0050] Figure 1 A schematic flow chart of a fatigue load suppression method in a wake wind farm provided in the present application;

[0051] Figure 2 A schematic flow chart of another method for suppressing fatigue loads in a wake wind farm provided in the present application;

[0052] Figure 3 A schematic diagram of the structure of an incoming wind speed prediction framework provided in this application;

[0053] Figure 4 A structural schematic diagram of a wake wind farm simulation system provided in this application;

[0054] FIG5 (a) is a comparison diagram of fatigue load suppression simulation under low wind speed provided by the present application;

[0055] FIG5( b ) is a comparison diagram of fatigue load suppression simulation under medium wind speed provided by the present application;

[0056] FIG5( c ) is a comparison diagram of fatigue load suppression simulation under high wind speed provided by the present application;

[0057] Figure 6 A wind speed prediction accuracy comparison chart provided in this application;

[0058] Figure 7 A schematic structural diagram of a fatigue load suppression device in a wake wind farm provided in the present application;

[0059] Figure 8 This is a schematic diagram of the structure of an electronic device provided in this application. DETAILED DESCRIPTION

[0060] The core of the present application is to provide a method for suppressing fatigue loads in a wake wind farm, which can suppress the fatigue loads in the wake wind farm more efficiently and accurately; another core of the present application is to provide a fatigue load suppression device, electronic device and computer-readable storage medium in a wake wind farm, all of which have the above-mentioned beneficial effects.

[0061] In order to describe the technical solutions in the embodiments of the present application more clearly and completely, the technical solutions in the embodiments of the present application will be introduced below in conjunction with the drawings 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. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0062] An embodiment of the present application provides a method for suppressing fatigue loads in a wake wind farm.

[0063] Please refer to Figure 1 , Figure 1 This is a schematic flow chart of a fatigue load suppression method in a wake wind farm provided in the present application. The fatigue load suppression method in a wake wind farm may include the following S101 to S106.

[0064] S101: Obtain historical wind speeds of target units in the wake wind farm, and use the historical wind speeds to train a wind speed prediction model.

[0065] S102: Processing the current wind speed of the target unit using a wind speed prediction model to obtain a predicted wind speed of the target unit.

[0066] The above steps are aimed at realizing model training based on historical wind speeds and future wind speed prediction based on the training model. In the implementation process, first, the historical wind speed data of the target unit in the wake wind farm (specifically, the most upstream unit) can be collected, and then the wind speed prediction model can be obtained based on these historical wind speed training. Obviously, the wind speed prediction model is used to realize future wind speed prediction; wherein, the model training can specifically be LS-SVM (least squares support vector machine) model training. Further, the current wind speed of the target unit is collected, and thus, it can be predicted using the wind speed prediction model to obtain the predicted wind speed in the future.

[0067] In one embodiment of the present application, the obtaining of the historical wind speed of the target unit in the wake wind farm may include: collecting a first preset number of historical wind speeds of the target unit in the wake wind farm at a preset time interval;

[0068] Accordingly, the above-mentioned use of the wind speed prediction model to process the current wind speed of the target unit to obtain the predicted wind speed of the target unit may include: using the wind speed prediction model to process the current wind speed of the target unit to obtain the predicted wind speed at each preset time interval in the prediction time domain.

[0069] It is understandable that in order to ensure the accuracy of the model and the wind speed prediction, the historical wind speed can be the wind speed data collected according to the preset frequency (that is, the preset time interval mentioned above), and correspondingly, the wind speed prediction based on the wind speed prediction model can also be a plurality of wind speed data predicted based on the preset frequency. It should be noted that the values ​​of the preset time interval, prediction time domain, and the number of historical wind speeds collected do not affect the implementation of the present technical solution, and can be set by technical personnel according to actual needs, and this application does not limit this. For example, the historical wind speed at intervals of 1 hour in the past day can be collected, so as to obtain the predicted wind speed at intervals of 1 hour in the future day based on the current wind speed prediction.

[0070] S103: Build a state space model of the wind turbine, and construct a dynamic wake model based on the predicted wind speed and the preset wake sensitivity coefficient.

[0071] This step aims to realize the construction of a dynamic wake model. It should be noted that the dynamic wake model proposed in this scheme simplifies the traditional nonlinear, coupled analytical wake model into a linear, decoupled model, which is suitable for fast wake calculation and can effectively ensure the calculation efficiency. In the implementation process, the state space model of the wind turbine (including the overall unit of the target unit) and the predicted wind speed are used to construct a dynamic wake model, and some wake sensitivity coefficients are taken into account during the construction process to effectively ensure the accuracy of the model, thereby effectively ensuring the accuracy of fatigue load suppression.

[0072] In a possible implementation, the preset wake sensitivity coefficient may include: a sensitivity coefficient of the wake wind speed of the wind turbine in the wind turbine group to the corresponding wind speed of the wind turbine; a sensitivity coefficient of the wake wind speed of the wind turbine in the wind turbine group to the corresponding wind speed of the wind turbine; a sensitivity coefficient of the wake wind speed of the wind turbine in the wind turbine group to the corresponding wind turbine pitch angle. It is understandable that these sensitivity coefficients can be calculated based on the basic parameters of the wind turbine group and its operating data.

[0073] In one embodiment of the present application, the state space model of the wind turbine generator set is constructed, and the dynamic wake model is constructed in combination with the predicted wind speed and the preset wake sensitivity coefficient, which may include:

[0074] Build a state space model of the wind turbine and determine the wind speed increment and mechanical state increment of the wind turbine in the wake wind farm based on the state space model;

[0075] A dynamic wake model is constructed based on the wind turbine wind speed increment, the wind turbine mechanical state increment, the predicted wind speed, and the preset wake sensitivity coefficient.

[0076] In the implementation process, the state space model of the wind turbine can be specifically a linearized two-degree-of-freedom model of the entire wake wind farm. By building the state space model of the wind turbine, the wind speed increment and the mechanical state increment of the wind turbine in the wake wind farm can be obtained, and then the dynamic wake model of the wake wind farm can be built by combining the predicted wind speed and the preset wake sensitivity coefficient. Furthermore, the state space model also includes the increment of the pitch angle control amount and the increment of the motor torque control amount, so as to realize the solution of the optimal (motor) torque and the optimal pitch angle in the subsequent calculation process.

[0077] Among them, the above-mentioned construction of the state space model of the wind turbine may include: constructing a transmission system model, a pitch angle control system model, and a motor torque response system model of the wind turbine according to various index parameters of the wind turbine; and using the transmission system model, the pitch angle control system model, and the motor torque response system model to construct the state space model of the wind turbine.

[0078] That is to say, the state space model of the wind turbine can be constructed based on the transmission system model, pitch angle control system model and motor torque response system model of the wind turbine, so as to achieve the solution of the optimal (motor) torque and the optimal pitch angle in the subsequent calculation process. Among them, each model can be calculated according to various index parameters of the wind turbine, such as the aerodynamic torque of the unit, the motor torque of the unit, air density, thrust coefficient, rotor speed, rotor mass, motor mass, pitch angle response time constant, motor torque response time constant, etc.

[0079] Among them, the above-mentioned construction of a dynamic wake model based on the wind turbine wind speed increment, the wind turbine mechanical state increment, the predicted wind speed, and the preset wake sensitivity coefficient may include: determining the time delay of wake propagation based on the unit spacing of the wind turbine units; and constructing a dynamic wake model based on wake time lag based on the wind turbine wind speed increment, the wind turbine mechanical state increment, the predicted wind speed, the preset wake sensitivity coefficient, and the time delay.

[0080] It is understandable that, depending on the distance between the units, the mechanical state of each unit and the change in wind speed generally require different time delays to be transmitted from the upstream unit to the downstream unit. Therefore, taking the time delay of wake propagation into account when constructing a dynamic wake model can further improve the model accuracy and ensure more accurate fatigue load suppression.

[0081] S104: Combine the dynamic wake model with the state space model to establish a discretized equation.

[0082] S105: Establishing a linearized expression for the tower bending moment and the shaft torque, and establishing an optimization objective function for the tower bending moment and the shaft torque based on the linearized expression and the discretized equation.

[0083] The above steps are aimed at achieving the construction of the optimization objective function. Specifically, the dynamic wake model of the wind turbine is combined with the state space model to establish a discretized equation. At the same time, the optimization objective function of the linear expression of the tower bending moment and shaft torque can be constructed by combining the linear expression of the tower bending moment and shaft torque. It can be understood that the tower bending moment and shaft torque are key indicators that affect the fatigue load of the wind turbine. By constructing the optimization objective function of the tower bending moment and shaft torque, it is aimed to effectively reduce the fluctuation of the two, so as to effectively reduce the fatigue load.

[0084] S106: Solving the optimization objective function under preset constraint conditions to obtain the optimal torque and the optimal pitch angle, and achieving fatigue load suppression according to the optimal torque and the optimal pitch angle.

[0085] This step aims to achieve fatigue load suppression based on optimal torque and optimal pitch angle. Specifically, the constraint conditions of the optimization objective function can be pre-constructed, and the optimization objective function can be solved under the preset constraint conditions to obtain the optimal torque and optimal pitch angle; further, the control instructions of the optimal torque and optimal pitch angle are generated and executed so that the wind turbine set operates according to the optimal torque and optimal pitch angle, thereby achieving fatigue load suppression of the wind turbine set.

[0086] In one possible manner, the above-mentioned preset constraints may include: the motor torque in the wind turbine set does not exceed the maximum motor torque; the motor pitch angle in the wind turbine set does not exceed the maximum motor pitch angle; the change in the motor torque in the wind turbine set satisfies the first preset range; the change in the motor pitch angle in the wind turbine set satisfies the second preset range.

[0087] It should be noted that the specific values ​​of the above-mentioned maximum motor torque, maximum motor pitch angle, first preset range and second preset range can be set by technical personnel in combination with the actual characteristics of the wind turbine generator set, and this application does not limit this.

[0088] It can be seen that the fatigue load suppression method in the wake wind farm provided in the embodiment of the present application uses the historical incoming wind speed in the wake wind farm to perform model training to obtain the predicted incoming wind speed in the future. The predicted wind speed can be used to estimate the future changes of the fatigue load and take action in advance to achieve fatigue load suppression; and this solution proposes fatigue load suppression based on a dynamic wake model, which simplifies the traditional nonlinear, coupled analytical wake model into a linear, decoupled model, which is suitable for fast wake calculation; combined with the internal dynamic characteristics of the wind turbine, it can predict the wake wind speed loss under the change of the mechanical state (generator torque and pitch angle) of the wind turbine in the future, thereby achieving fatigue load suppression on the tower and blades. It can be seen that the technical solution achieves more efficient and accurate fatigue load suppression in the wake wind farm.

[0089] Based on this, please refer to Figure 2 , Figure 2 A schematic flow chart of another method for suppressing fatigue loads in a wake wind farm provided in the present application, the implementation process of which may include:

[0090] 1. Collect the wind speed of the upstream unit in the wake wind farm The historical data is used as the training wind speed, and the wind speed collection frequency is set for each prediction interval. Collect a wind speed value.

[0091] 2. LS-SVM training is performed based on the historical data of wind speed of the most upstream unit. The calculation expression of LS-SVM can be expressed as:

[0092] (1);

[0093] in, , is a constant, is the total number of historical wind speeds, is the kernel function, in which represents the output vector, Represents the input vector. Since the radial basis function (RBF) kernel can handle nonlinear relationships well and has good comprehensive performance, the RBF kernel can be selected as the kernel function, and its expression is:

[0094] (2);

[0095] in, Represented as kernel function width.

[0096] During the training process, The historical wind speed samples are divided into multiple segments, each segment contains The training input of historical wind speed and the output of training are The wind speed after the time is the training output. Taking time as an example, the specific process of wind speed training is as follows:

[0097] (3).

[0098] 3. Please refer to Figure 3 , Figure 3 This is a schematic diagram of the structure of the wind speed prediction framework provided by this application. The overall process is divided into two parts: training and prediction. After the training is completed, the current wind speed of the upstream unit is obtained according to the LS-SVM model obtained by training and the measured wind speed. , for future prediction time domain Each prediction interval The predicted wind speed is used for subsequent optimization of fatigue loads. Indicates the current moment. The specific process is:

[0099] (4);

[0100] …;

[0101] (5).

[0102] 4. By constructing the wake sensitivity coefficient, the traditional Jensen engineering wake model is linearized so that the wake model can take into account both prediction accuracy and calculation speed. The specific calculation process of the wake sensitivity coefficient is as follows:

[0103] (6);

[0104] (7);

[0105] (8);

[0106] in, , Represents any two wind turbines in the wind farm. When i=j, i and j represent the same wind turbine, and n is the total number of wind turbines in the wind farm. represents the sensitivity coefficient of the wake wind speed of wind turbine i to the wind speed of wind turbine j; It is expressed as the sensitivity coefficient of the wake wind speed of fan i to the speed of fan j; It is expressed as the sensitivity coefficient of the wake wind speed of wind turbine i to the pitch angle of wind turbine j. represents the radius of fan j, represents the radius of fan i, is the wake expansion radius of fan j’s wake at fan i, is the wake expansion coefficient, is the distance between the upstream fan j and the downstream fan i, Is the fan in The thrust coefficient at time, is the wake coverage area of ​​the upstream unit j to the downstream unit i, is the rotor swept area of ​​fan i, For the fan Measure wind speed at any time, is the thrust coefficient of fan j About Tip Speed ​​Ratio The sensitivity coefficient, is the thrust coefficient of fan j About pitch angle The sensitivity coefficient of .

[0107] 5. Construct the state space model of the wind turbine. The specific process is as follows:

[0108] The transmission system of a wind turbine can be expressed as:

[0109] (9);

[0110] (10);

[0111] (11);

[0112] in, is the aerodynamic torque of the unit, is the motor torque of the unit, is the air density, is the thrust coefficient, is the rotor speed, The rotor mass and motor quality The equivalent mass obtained is calculated as , is the gearbox ratio.

[0113] The pitch angle control system can be modeled as a first-order inertial model:

[0114] (12);

[0115] in, is the pitch angle control command, is the pitch angle response time constant.

[0116] The response process of the motor torque can also be modeled as a first-order inertia model:

[0117] (13);

[0118] in, is the motor torque control command, is the motor torque response time constant.

[0119] According to formulas (9), (10), and (11), at the current moment Performing Taylor expansion, the partial derivative of the motor speed can be expressed as:

[0120] (14);

[0121] in, is the pneumatic torque increment, is the motor torque increment, for The aerodynamic torque at the moment, for The motor torque at the moment.

[0122] According to formula (12), we can get:

[0123] (15);

[0124] According to formula (13), we can get:

[0125] (16);

[0126] Combining formulas (14), (15), and (16), the linearized two-degree-of-freedom model of the entire wake wind farm can be described as:

[0127] (17);

[0128] ;

[0129] ;

[0130] ;

[0131] ;

[0132] ;

[0133] ;

[0134] ;

[0135] in, is the increment of the pitch angle control amount, is the increment of the motor torque control quantity, is the vector of wind speed increments of all wind turbines in the wind farm, is a vector composed of the mechanical state increments of all wind turbines in the wind farm, It is a vector composed of all wind turbine control variables in the wind farm.

[0136] Taking the nth fan as an example, , , and The matrix is ​​specifically:

[0137] ; ; ; ;

[0138] in, , and They represent the sensitivity coefficients of the aerodynamic torque of the wind turbine n to its own speed, pitch angle and wind speed respectively.

[0139] The nth fan , , and As an example, the , , and matrix.

[0140] 6. Wind farm predicted in step 5 , and the wind speed of the upstream unit Construct a dynamic wake model. The specific process is as follows:

[0141] (18);

[0142] in, is the wind speed change vector of the upstream unit in the wind farm, and its calculation expression is:

[0143] ;

[0144] and is the wake sensitivity coefficient matrix, which is composed of the wake sensitivity coefficients described by equations (6), (7), and (8):

[0145] ; ;

[0146] in, is the predicted wind speed of each unit, is the vector composed of wind speeds of all wind turbines in the wind farm, .

[0147] 7. According to the distance between units , determine the time delay of wake propagation :

[0148] (19).

[0149] 8. Establish a dynamic wake model considering wake time lag.

[0150] At the current moment For example, according to the different distances between the units, the mechanical status of each unit and wind speed The changes require different time lags Only then can it be transmitted to the downstream units it affects. The specific process is as follows:

[0151] (20);

[0152] in, is the mechanical time lag, It is the time lag of the incoming wind speed of the most upstream unit.

[0153] 9. According to the derived wind turbine model and dynamic wake model, the discretization equation is established. The specific process is as follows:

[0154] (twenty one);

[0155] in, ; ; ; ; is the sampling time.

[0156] In model predictive control, by Set a prediction horizon to describe the optimization problem , by expanding formula (21) into Moving forward in time, we get:

[0157] (twenty two);

[0158] Where h is the total prediction step, and its value is h = / . is Moment prediction The mechanical state of the moment, is Moment prediction Control actions at all times, yes The predicted wind speed reduction value at the time.

[0159] 10. Take tower bending moment and shaft torque as key indicators affecting the fatigue load of wind turbines, and reduce fatigue load by reducing their fluctuations. The calculation expressions of tower bending moment and shaft torque are:

[0160] (twenty three);

[0161] (twenty four);

[0162] (25);

[0163] in, It's thrust. is the tower height, is the tower bending moment, is the shaft torque.

[0164] 11. Establish the linearized expressions of tower bending moment and shaft torque. The specific process is as follows:

[0165] (26);

[0166] (27);

[0167] in, , , , Respectively expressed as the tower bending moment and shaft torque of each unit with respect to the mechanical state and wind speed The sensitivity coefficient matrix is: is the vector composed of the bending moment increments of each wind turbine tower, is the vector composed of the torque increments of each fan shaft.

[0168] 12. According to formulas (22), (26), and (27), the optimization objective function of tower bending moment and shaft torque based on model predictive control is established to reduce the fluctuation of the two. The specific process is as follows:

[0169] (28);

[0170] in, represents the optimization objective function, The weight coefficient for suppressing tower bending moment fluctuations is: is the weight coefficient for suppressing shaft torque fluctuation.

[0171] 13. The constraints of motor torque and pitch angle are:

[0172] (29);

[0173] in, and is the maximum value of motor torque and pitch angle, , , and They are the maximum and minimum changes of motor torque and pitch angle respectively.

[0174] 14. Under the constraints of formula (29), solve the optimization objective function (28) to obtain the optimal torque and pitch angle The control commands are used to achieve the optimal fatigue load control of the wake wind farm.

[0175] Furthermore, in order to verify the effectiveness of the above-mentioned fatigue load suppression method for the wind speed prediction of the wake wind farm, a Figure 4 The typical wake wind farm simulation system shown in Figure 4A structural schematic diagram of a wake wind farm simulation system provided in this application. Considering the 12×5 MW wind turbine arrangement under study, the wind turbine model used is the NREL 5 MW wind turbine, and each wind turbine is connected to the feeder through a 0.9 / 35 kV transformer. The wind turbines are placed in four rows and three columns, with a row spacing of five rotor diameters (5D), a column spacing of 5D, and a rotor diameter D of 126 m. All wind turbines are collected at the connecting line and transmitted to the external power grid through a 155 / 380 kV main transformer. It is assumed that the incoming wind speed blows out in the longitudinal direction. On this basis, in order to compare the fatigue load suppression effect under different wind speeds, low, medium and high wind speeds can be used as simulation conditions respectively. Figure 5 (a) is a fatigue load suppression simulation comparison diagram under low wind speed provided by the present application. The wind speed range is (7m / s-9m / s). It can be seen that the proposed method has a smaller amplitude and number of cycles, which means that the wind machine will generate less fatigue load; Figure 5 (b) is a fatigue load suppression simulation comparison diagram under medium wind speed provided by the present application. The wind speed range is (9m / s-12m / s). It can be seen that the proposed method will generate less fatigue load; Figure 5 (c) is a fatigue load suppression simulation comparison diagram under high wind speed provided by the present application. The wind speed range is (12m / s-14m / s). It can be seen that the proposed method also has a good fatigue load suppression effect under high wind speed. In addition, please refer to Table 1, which is a fatigue load suppression effect information table provided in this application. Taking wind turbine 2 as the research object, it can be seen from Table 1 that the proposed control method has the best performance of reducing the equivalent damage load of Mt (7.753MNm) and Ts (0.072MNm) at low wind speed. At medium wind speed, the proposed control method also has the best effect of reducing fatigue load, Mt (5.918MNm) and Ts (0.113MNm); at high wind speed, the optimization effect of fatigue load is Mt (7.365MNm) and Ts (0.233MNm).

[0176] Table 1 Information table of the suppression effect of a fatigue load

[0177]

[0178] From the simulation comparison results of the above three wind speed ranges, it can be seen that the proposed wake wind farm wind speed prediction fatigue load suppression strategy can predict the change of wind speed and take action in advance to control the motor torque and pitch angle reference value, so that the wind turbine can respond to complex wind speed fluctuations in time, reduce the fluctuation of the mechanical load of the unit, and effectively reduce the fatigue load of the wake wind farm.

[0179] Furthermore, to verify the prediction accuracy of the incoming wind speed of the proposed LS-SVM strategy, please refer to Figure 6 , Figure 6 A wind speed prediction accuracy comparison chart provided in this application, from Figure 6 It can be seen that both methods have high wind speed prediction accuracy, but the wind speed predicted by LS-SVM is closer to the characteristics of the actual wind speed.

[0180] In addition, in order to evaluate the wind speed prediction performance, the root mean square error (RMSE) and determination coefficient (R2) of the proposed scheme were calculated, and the results are shown in Table 2, which is a wind speed prediction accuracy information table provided in this application. As can be seen from Table 2, the RMSE of the proposed LS-SVM method, SVM method and LSTM method are approximately 0.0286, 0.0685 and 0.0491, respectively. The value of LS-SVM is the smallest, indicating that the algorithm has good prediction performance. The R2 values ​​of LS-SVM, SVM and LSTM are 0.9956, 0.9479 and 0.9884, respectively. The closer the value of R2 is to 1, the better the model fits the wind speed. Therefore, the adopted LS-SVM method has better wind speed prediction accuracy.

[0181] Table 2 A wind speed prediction accuracy information table

[0182]

[0183] It can be seen that the fatigue load suppression method in the wake wind farm provided by the embodiment of the present application has the following advantages:

[0184] (1) The least squares support vector machine (LS-SVM) is used to train the LS-SVM based on the historical incoming wind speed of the wake wind farm, and the predicted incoming wind speed value for each prediction interval in the prediction time domain can be obtained. The predicted wind speed can be used to estimate the future changes of fatigue loads and take action in advance to suppress fatigue loads. Compared with the Lidar-based prediction method, the proposed method can reduce the investment cost of wake wind farm equipment.

[0185] (2) The dynamic wake model is used to simplify the traditional nonlinear, coupled analytical wake model into a linear, decoupled model, which is suitable for fast wake calculation. Combined with the internal dynamic characteristics of the wind turbine, it can predict the wake wind speed loss under the future changes in the mechanical state of the wind turbine (such as generator torque and pitch angle). The wake wind speed predicted by the dynamic wake model can be used to dispatch the mechanical state and suppress the fatigue load on the tower and blades.

[0186] (3) A fatigue load suppression control scheme based on model predictive control is proposed, which realizes wake prediction control and tracks the wake wind farm power command demand. It takes into account the coupling relationship between active power, wind speed and mechanical load, and can achieve more accurate load prediction control and more effective fatigue load suppression. In addition, the proposed wake optimization method based on model predictive control has low computational complexity, reduces the computational burden of the central controller, and allows real-time control.

[0187] An embodiment of the present application provides a fatigue load suppression device in a wake wind farm.

[0188] Please refer to Figure 7 , Figure 7 This is a schematic structural diagram of a fatigue load suppression device in a wake wind farm provided in the present application. The fatigue load suppression device in the wake wind farm may include:

[0189] Training module 1 is used to obtain the historical wind speed of the target unit in the wake wind farm and obtain a wind speed prediction model by using the historical wind speed training;

[0190] Prediction module 2, used to process the current wind speed of the target unit using the wind speed prediction model to obtain the predicted wind speed of the target unit;

[0191] The first building module 3 is used to build a state space model of the wind turbine and construct a dynamic wake model in combination with the predicted wind speed and the preset wake sensitivity coefficient;

[0192] The second building module 4 is used to combine the dynamic wake model with the state space model to establish a discretized equation;

[0193] The third establishment module 5 is used to establish a linearized expression about the tower bending moment and the shaft torque, and establish an optimization objective function about the tower bending moment and the shaft torque according to the linearized expression and the discretized equation;

[0194] The solution module 6 is used to solve the optimization objective function under preset constraints to obtain the optimal torque and the optimal pitch angle, and to achieve fatigue load suppression according to the optimal torque and the optimal pitch angle.

[0195] It can be seen that the fatigue load suppression device in the wake wind farm provided by the embodiment of the present application performs model training through the historical incoming wind speed in the wake wind farm to obtain the predicted value of the incoming wind speed in the future time. The predicted wind speed can be used to estimate the future changes of the fatigue load and take action in advance to achieve the suppression of the fatigue load; and the present solution proposes fatigue load suppression based on the dynamic wake model, which simplifies the traditional nonlinear, coupled analytical wake model into a linear, decoupled model, which is suitable for fast wake calculation; combined with the internal dynamic characteristics of the wind turbine, it can predict the wake wind speed loss under the change of the mechanical state (generator torque and pitch angle) of the wind turbine in the future, and then achieve fatigue load suppression on the tower and blades. It can be seen that the present technical solution achieves more efficient and accurate fatigue load suppression in the wake wind farm.

[0196] In one embodiment of the present application, the training module 1 may be specifically used to collect a first preset number of historical wind speeds of the target unit in the wake wind farm at a preset time interval;

[0197] Accordingly, the prediction module 2 can be specifically used to process the current wind speed of the target unit using the wind speed prediction model to obtain the predicted wind speed at each preset time interval within the prediction time domain.

[0198] In one embodiment of the present application, the first establishing module 3 may include:

[0199] A determination unit is used to build a state space model of the wind turbine and determine the wind speed increment and the mechanical state increment of the wind turbine in the wake wind farm according to the state space model;

[0200] The construction unit is used to construct a dynamic wake model according to the wind speed increment of the wind turbine, the mechanical state increment of the wind turbine, the predicted wind speed, and the preset wake sensitivity coefficient.

[0201] In one embodiment of the present application, the above-mentioned determination unit can be specifically used to construct a transmission system model, a pitch angle control system model, and a motor torque response system model of the wind turbine according to various index parameters of the wind turbine; and to construct a state space model of the wind turbine using the transmission system model, the pitch angle control system model, and the motor torque response system model.

[0202] In one embodiment of the present application, the above-mentioned construction unit can be specifically used to determine the time delay of wake propagation according to the unit spacing of the wind turbine set; and to construct a dynamic wake model based on wake time lag according to the wind turbine wind speed increment, the wind turbine mechanical state increment, the predicted wind speed, the preset wake sensitivity coefficient, and the time delay.

[0203] In one embodiment of the present application, the preset wake sensitivity coefficient may include:

[0204] The sensitivity coefficient of the wind turbine wake wind speed in the wind turbine group to the corresponding wind turbine wind speed;

[0205] The sensitivity coefficient of the wind turbine wake wind speed in the wind turbine to the corresponding wind turbine speed;

[0206] The sensitivity coefficient of the wind turbine wake wind speed in a wind turbine to the corresponding wind turbine pitch angle.

[0207] In one embodiment of the present application, the above-mentioned preset constraint conditions may include:

[0208] The motor torque in the wind turbine does not exceed the maximum motor torque;

[0209] The motor pitch angle in the wind turbine does not exceed the maximum motor pitch angle;

[0210] The change in the motor torque in the wind turbine generator set satisfies a first preset range;

[0211] The variation of the motor pitch angle in the wind turbine generator set satisfies the second preset range.

[0212] For an introduction to the device provided in the embodiments of the present application, please refer to the above method embodiments, and the present application will not elaborate on them here.

[0213] An embodiment of the present application provides an electronic device.

[0214] Please refer to Figure 8 , Figure 8 This is a schematic diagram of the structure of an electronic device provided in this application, and the electronic device may include:

[0215] A memory 11, used for storing computer programs;

[0216] The processor 10 can implement the steps of any one of the above-mentioned fatigue load suppression methods in a wake wind farm when executing the computer program.

[0217] like Figure 8 , which is a schematic diagram of the composition structure of an electronic device, the electronic device may include: a processor 10, a memory 11, a communication interface 12 and a communication bus 13. The processor 10, the memory 11 and the communication interface 12 all communicate with each other through the communication bus 13.

[0218] In the embodiment of the present application, the processor 10 may be a central processing unit (CPU), an application specific integrated circuit, a digital signal processor, a field programmable gate array or other programmable logic devices, etc.

[0219] The processor 10 may call a program stored in the memory 11 . Specifically, the processor 10 may execute operations in an embodiment of the fatigue load suppression method in a wake wind farm.

[0220] The memory 11 is used to store one or more programs, which may include program codes, and the program codes include computer operation instructions. In the embodiment of the present application, the memory 11 at least stores programs for implementing the following functions:

[0221] Obtain the historical wind speed of the target unit in the wake wind farm, and use the historical wind speed to train a wind speed prediction model;

[0222] The current wind speed of the target unit is processed using the wind speed prediction model to obtain the predicted wind speed of the target unit;

[0223] Build a state space model of the wind turbine and construct a dynamic wake model based on the predicted wind speed and the preset wake sensitivity coefficient;

[0224] The dynamic wake model is combined with the state space model to establish the discretized equation;

[0225] Establishing linearized expressions for tower bending moment and shaft torque, and establishing optimization objective functions for tower bending moment and shaft torque based on the linearized expressions and discretized equations;

[0226] The optimization objective function is solved under preset constraints to obtain the optimal torque and the optimal pitch angle, and fatigue load suppression is achieved based on the optimal torque and the optimal pitch angle.

[0227] In a possible implementation, the memory 11 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and an application required for at least one function, etc.; the data storage area may store data created during use.

[0228] In addition, the memory 11 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device or other volatile solid-state storage device.

[0229] The communication interface 12 may be an interface of a communication module, and is used to connect to other devices or systems.

[0230] Of course, it should be noted that Figure 8 The structure shown does not constitute a limitation on the electronic device in the embodiment of the present application. In actual applications, the electronic device may include Figure 8 More or fewer components than shown, or combinations of certain components.

[0231] An embodiment of the present application provides a computer-readable storage medium.

[0232] A computer program is stored on a computer-readable storage medium provided in an embodiment of the present application. When the computer program is executed by a processor, the steps of any one of the above-mentioned methods for suppressing fatigue loads in a wake wind farm can be implemented.

[0233] The computer-readable storage medium may include: a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0234] For an introduction to the computer-readable storage medium provided in the embodiments of the present application, please refer to the above method embodiments, and the present application will not elaborate on them here.

[0235] The various embodiments in the specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0236] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0237] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0238] The technical solution provided by the present application is described in detail above. The principle and implementation method of the present application are described in detail using specific examples herein, and the description of the above embodiments is only used to help understand the method and core idea of ​​the present application. It should be pointed out that for ordinary technicians in this technical field, without departing from the principle of the present application, several improvements and modifications can be made to the present application, and these improvements and modifications also fall within the scope of protection of the present application.

Claims

1. A method for suppressing fatigue load in a wake wind farm, characterized in that: include: Obtaining historical wind speeds of target units in the wake wind farm, and using the historical wind speeds to train a wind speed prediction model; Processing the current wind speed of the target unit using the wind speed prediction model to obtain the predicted wind speed of the target unit; Building a state space model of the wind turbine, and building a dynamic wake model in combination with the predicted wind speed and the preset wake sensitivity coefficient; Combining the dynamic wake model with the state space model to establish a discretized equation; Establishing a linearized expression for tower bending moment and shaft torque, and establishing an optimization objective function for tower bending moment and shaft torque based on the linearized expression and the discretized equation; The optimization objective function is solved under preset constraint conditions to obtain an optimal torque and an optimal pitch angle, and fatigue load suppression is achieved based on the optimal torque and the optimal pitch angle.

2. The fatigue load suppression method in a wake wind farm according to claim 1, characterized in that: Obtain the historical wind speed of the target unit in the wake wind farm, including: At a preset time interval, collecting a first preset number of the historical wind speeds of the target unit in the wake wind farm; Accordingly, the current wind speed of the target unit is processed by using the wind speed prediction model to obtain the predicted wind speed of the target unit, including: The wind speed prediction model is used to process the current wind speed of the target unit to obtain the predicted wind speed at each preset time interval within the prediction time domain.

3. The fatigue load suppression method in a wake wind farm according to claim 1, characterized in that: A state space model of the wind turbine is constructed, and a dynamic wake model is constructed in combination with the predicted wind speed and the preset wake sensitivity coefficient, including: Building a state space model of the wind turbine set, and determining a wind speed increment and a mechanical state increment of a wind turbine in the wake wind farm according to the state space model; The dynamic wake model is constructed according to the wind speed increment of the wind turbine, the mechanical state increment of the wind turbine, the predicted wind speed, and the preset wake sensitivity coefficient.

4. The fatigue load suppression method in a wake wind farm according to claim 3, characterized in that: Build a state space model of the wind turbine, including: Constructing a transmission system model, a pitch angle control system model, and a motor torque response system model of the wind turbine generator set according to various index parameters of the wind turbine generator set; The state space model of the wind turbine generator set is constructed by using the transmission system model, the pitch angle control system model and the motor torque response system model.

5. The fatigue load suppression method in a wake wind farm according to claim 3, characterized in that: The dynamic wake model is constructed according to the wind speed increment of the wind turbine, the mechanical state increment of the wind turbine, the predicted wind speed, and the preset wake sensitivity coefficient, including: Determining a time delay for wake propagation based on the unit spacing of the wind turbine units; A dynamic wake model based on wake time lag is constructed according to the wind speed increment of the wind turbine, the mechanical state increment of the wind turbine, the predicted wind speed, the preset wake sensitivity coefficient, and the time delay.

6. The fatigue load suppression method in a wake wind farm according to claim 1, characterized in that: The preset wake sensitivity coefficient includes: The sensitivity coefficient of the wind turbine wake wind speed in the wind turbine group to the corresponding wind turbine wind speed; The sensitivity coefficient of the wind turbine wake wind speed in the wind turbine generator set to the corresponding wind turbine speed; The sensitivity coefficient of the wind turbine wake wind speed in the wind turbine generator set to the corresponding wind turbine pitch angle.

7. The fatigue load suppression method in a wake wind farm according to claim 1, characterized in that: The preset constraints include: The motor torque in the wind turbine generator set does not exceed the maximum motor torque; The motor pitch angle in the wind turbine set does not exceed the maximum motor pitch angle; The change in the motor torque in the wind turbine generator set satisfies a first preset range; The variation of the motor pitch angle in the wind turbine generator set satisfies a second preset range.

8. A fatigue load suppression device in a wake wind farm, characterized in that: include: A training module, used to obtain the historical wind speed of the target unit in the wake wind farm, and use the historical wind speed to train a wind speed prediction model; A prediction module, used to process the current wind speed of the target unit using the wind speed prediction model to obtain the predicted wind speed of the target unit; A first building module is used to build a state space model of the wind turbine and to construct a dynamic wake model in combination with the predicted wind speed and the preset wake sensitivity coefficient; A second building module is used to combine the dynamic wake model with the state space model to establish a discretized equation; A third establishing module is used to establish a linearized expression about the tower bending moment and the shaft torque, and to establish an optimization objective function about the tower bending moment and the shaft torque according to the linearized expression and the discretized equation; A solution module is used to solve the optimization objective function under preset constraints to obtain an optimal torque and an optimal pitch angle, and to achieve fatigue load suppression according to the optimal torque and the optimal pitch angle.

9. An electronic device, characterized in that: include: Memory for storing computer programs; A processor is used to implement the steps of the fatigue load suppression method in a wake wind farm as claimed in any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the fatigue load suppression method in a wake wind farm according to any one of claims 1 to 7 are implemented.

Images