Wind turbine generator aerodynamic performance real-time optimization method and device, medium and equipment

Abstract

The invention relates to the technical field of wind power generation, and provides a wind turbine generator aerodynamic performance real-time optimization method and device, a medium and equipment, and the method comprises the steps: collecting the sensing data of a target wind turbine generator group based on a sensor preset in the target wind turbine generator group; taking the maximum total power generation efficiency of the target wind turbine generator group and the minimum sum of the accumulated damage degrees of all the target wind turbine generators as optimization objectives to construct a dual-objective fitness function; solving the optimal pitch angle and yaw angle instruction set of the unit group based on the dual-target fitness function, the environmental data, the operating parameters of each target wind turbine generator and a preset genetic algorithm; and the optimal pitch angle and yaw angle instruction corresponding to each target wind turbine generator is issued to a terminal layer of the corresponding target wind turbine generator to be executed, so that the cooperative action of each target wind turbine generator in the target wind turbine generator group is realized. The overall power generation efficiency of the wind turbine generator group can be effectively improved, meanwhile, the operation damage degree of the wind turbine generators is reduced, and the service life of the wind turbine generators is prolonged.

Description

Technical Field

[0001] This disclosure relates to the field of wind power generation technology, and more specifically, to a method, apparatus, medium, and equipment for real-time optimization of the aerodynamic performance of a wind turbine. Background Technology

[0002] Against the backdrop of an accelerated shift in the energy structure towards clean energy, wind power, as a key component of the renewable energy sector, has seen its installed capacity and power generation scale continuously increase. As the core equipment of wind power generation, the operating efficiency and reliability of wind turbine units directly affect the economic benefits and energy output stability of the entire wind farm.

[0003] However, wind farms often consist of multiple wind turbines, which have complex dynamic coupling effects. Furthermore, the wake effect significantly reduces the overall power generation efficiency. In addition, wind turbines operate in a complex and variable natural environment, with key components such as blades and gearboxes subjected to alternating loads, leading to accumulated fatigue damage and a gradual shortening of their remaining lifespan.

[0004] In related technologies, the operation and maintenance (O&M) and control of wind turbine generators mainly rely on SCADA systems. However, SCADA systems cannot model the dynamic coupling effects between units, making it difficult to accurately grasp the mutual influence relationships between units, which may lead to inaccurate predictions of wind turbine fatigue losses. Furthermore, fatigue life assessments of wind turbine generators mostly rely on periodic inspections, failing to monitor the remaining lifespan of key components in real time, resulting in a lack of timeliness and specificity in O&M decisions. Summary of the Invention

[0005] This disclosure provides at least one method, apparatus, medium, and equipment for real-time optimization of the aerodynamic performance of wind turbine generator sets, which can effectively improve the overall power generation efficiency of wind turbine generator sets, while reducing the degree of operational damage to wind turbine generator sets and extending their service life.

[0006] This disclosure provides a method for real-time optimization of the aerodynamic performance of a wind turbine, including:

[0007] Sensor data about the target wind turbine group is collected based on sensors pre-installed in the target wind turbine group; wherein, the sensor data includes environmental data and operating parameters of each target wind turbine group;

[0008] The optimization objectives are to maximize the total power generation efficiency of the target wind turbine group and minimize the sum of the cumulative damage of all target wind turbines, and a dual-objective fitness function is constructed.

[0009] Based on the bi-objective fitness function, the environmental data, the operating parameters of each target wind turbine, and the preset genetic algorithm, the optimal pitch angle and yaw angle command set of the target wind turbine group is solved; wherein, the optimal pitch angle and yaw angle command set includes the optimal pitch angle and yaw angle command corresponding to each target wind turbine.

[0010] The optimal pitch angle and yaw angle commands corresponding to each target wind turbine are sent to the terminal layer of the corresponding target wind turbine for execution, so as to realize the coordinated operation of each target wind turbine in the target wind turbine group.

[0011] This disclosure provides a real-time aerodynamic performance optimization device for wind turbine generators, comprising:

[0012] The data acquisition module is used to acquire sensing data about the target wind turbine group based on sensors pre-installed in the target wind turbine group; wherein, the sensing data includes environmental data and operating parameters of each target wind turbine;

[0013] The function construction module is used to construct a dual-objective fitness function with the optimization objectives of maximizing the total power generation efficiency of the target wind turbine group and minimizing the sum of the cumulative damage of all target wind turbines.

[0014] The instruction solving module is used to solve for the optimal pitch angle and yaw angle instruction set of the target wind turbine group based on the dual-objective fitness function, the environmental data, the operating parameters of each target wind turbine, and a preset genetic algorithm; wherein, the optimal pitch angle and yaw angle instruction set includes the optimal pitch angle and yaw angle instructions corresponding to each target wind turbine.

[0015] The instruction issuing module is used to issue the optimal pitch angle and yaw angle instructions corresponding to each target wind turbine to the terminal layer of the corresponding target wind turbine for execution, so as to realize the coordinated action of each target wind turbine in the target wind turbine group.

[0016] This disclosure provides a computer device, including a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the computer device is running, the processor communicates with the memory via the bus. When the machine-readable instructions are executed by the processor, the real-time aerodynamic performance optimization method for wind turbines as described in any of the above possible embodiments is executed.

[0017] This disclosure provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the real-time aerodynamic performance optimization method for wind turbines as described in any of the above possible embodiments.

[0018] The real-time aerodynamic performance optimization method, apparatus, medium, and equipment for wind turbines provided in this disclosure are based on sensor data collected from a target wind turbine group by sensors pre-installed on the target wind turbine group. This sensor data includes environmental data and operating parameters of each target wind turbine. A dual-objective fitness function is constructed with the optimization objectives of maximizing the total power generation efficiency of the target wind turbine group and minimizing the sum of cumulative damage of all target wind turbines. This function comprehensively considers power generation efficiency and equipment safety, and provides a basis for subsequent optimization by quantitatively evaluating the power generation efficiency and turbine damage under different combinations of pitch and yaw angles. Based on the dual-objective fitness function, environmental data, operating parameters of each target wind turbine, and a pre-set genetic algorithm, the optimal pitch and yaw angle command set for the target wind turbine group is solved. The optimal pitch and yaw angle command set includes the optimal pitch and yaw angle commands corresponding to each target wind turbine. The genetic algorithm searches for the optimal solution in the solution space by simulating a natural evolutionary process, ensuring that the command set meets the dual-objective optimization requirements. The optimal pitch angle and yaw angle commands corresponding to each target wind turbine are sent to the terminal layer of the corresponding target wind turbine for execution, so as to realize the coordinated action of each target wind turbine in the target wind turbine group.

[0019] In this way, by constructing a dual-objective fitness function and combining it with a genetic algorithm to solve for the optimal instruction set, the pitch angle and yaw angle of each unit are uniformly adjusted, thereby realizing the real-time optimization of the aerodynamic performance of the wind turbine units. This effectively improves the overall power generation efficiency of the wind turbine group, while reducing the cumulative damage of the units and extending the service life of the equipment.

[0020] To make the above-mentioned objects, features and advantages of this disclosure more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of this disclosure, the accompanying drawings referenced in the embodiments will be briefly described below. These drawings are incorporated in and constitute a part of this specification. They illustrate embodiments conforming to this disclosure and, together with the specification, serve to explain the technical solutions of this disclosure. It should be understood that the following drawings only show some embodiments of this disclosure and should not be considered as limiting the scope. Those skilled in the art can obtain other related drawings based on these drawings without creative effort.

[0022] Figure 1 A flowchart of a real-time optimization method for the aerodynamic performance of a wind turbine provided in an embodiment of this disclosure is shown;

[0023] Figure 2A flowchart is shown below illustrating a method for solving the optimal pitch angle and yaw angle command set provided in an embodiment of this disclosure.

[0024] Figure 3 A flowchart illustrating a method for calculating the cumulative damage of a target wind turbine provided in an embodiment of this disclosure is shown.

[0025] Figure 4 A flowchart of a command population evolution method provided by an embodiment of this disclosure is shown;

[0026] Figure 5 A flowchart of a wind turbine operation status monitoring and maintenance early warning method provided in an embodiment of this disclosure is shown;

[0027] Figure 6 This diagram illustrates the structure of a real-time aerodynamic performance optimization device for wind turbine generators provided in an embodiment of this disclosure.

[0028] Figure 7 A schematic diagram of the structure of a computer device provided in an embodiment of this disclosure is shown. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of the embodiments of this disclosure clearer, the technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this disclosure, and not all of them. The components of the embodiments of this disclosure described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this disclosure provided in the accompanying drawings is not intended to limit the scope of the claimed disclosure, but merely represents selected embodiments of this disclosure. All other embodiments obtained by those skilled in the art based on the embodiments of this disclosure without inventive effort are within the scope of protection of this disclosure.

[0030] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0031] In this document, the term "and / or" merely describes a relationship, indicating that three relationships can exist. For example, A and / or B can represent three cases: A alone, A and B simultaneously, and B alone. Furthermore, the term "at least one" in this document means any combination of at least two of any one or more elements. For example, including at least one of A, B, and C can mean including any one or more elements selected from the set consisting of A, B, and C.

[0032] To facilitate understanding of this embodiment, the executing entity of the real-time aerodynamic performance optimization method for wind turbines provided in this disclosure will first be described in detail. The executing entity of the real-time aerodynamic performance optimization method for wind turbines provided in this disclosure is a computer device. This computer device can be a server. Specifically, the server can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud storage, big data, and artificial intelligence platforms.

[0033] The method for real-time optimization of the aerodynamic performance of wind turbine generators provided in this application will be described in detail below with reference to the accompanying drawings. See also Figure 1 The diagram shows a flowchart of a real-time aerodynamic performance optimization method for wind turbine generators provided in this embodiment of the present disclosure. The method includes the following steps S101 to S104:

[0034] S101, collecting sensing data about the target wind turbine group based on sensors pre-installed in the target wind turbine group.

[0035] It is understandable that a target wind turbine cluster includes multiple target wind turbines, which may be located within the same wind farm and have certain correlations in terms of geographical location and operating environment. A sensor is a device or apparatus that can sense a specified measurand and convert it into a usable signal according to a certain rule; it typically includes various types. For example, wind speed sensors utilize the principles of wind cups or hot-wire anemometers to convert the physical quantity of wind speed into an electrical signal, used to measure the wind speed in the environment where the wind turbine is located; wind direction sensors often use a wind vane structure, converting the wind direction angle into an electrical signal through photoelectric encoders or potentiometers to determine the direction of the wind; temperature sensors, such as thermocouples or resistance temperature detectors (RTDs), can convert temperature changes into electrical signals to monitor ambient temperature and the temperature of key components of the wind turbine; vibration sensors generally use accelerometers to detect the vibration of the wind turbine during operation, converting it into an electrical signal to analyze whether the unit is operating normally; speed sensors often utilize the Hall effect or photoelectric encoding principles to measure the rotational speed of the wind turbine blades or generator and convert the speed information into an electrical signal.

[0036] Here, the sensor data collected by various sensors pre-installed in the target wind turbine group can include environmental data and the operating parameters of each target wind turbine. Environmental data can include wind speed, wind direction, temperature, and air pressure, reflecting the external natural environmental conditions of the wind turbines and directly affecting their aerodynamic performance. For example, wind speed determines the wind force on the turbine blades, thus affecting power generation; wind direction determines the angle at which the blades receive the wind, affecting aerodynamic efficiency. The operating parameters of each target wind turbine can include its rotational speed, power, and vibration levels, reflecting the turbine's own operating status. For example, rotational speed reflects the efficiency of the wind turbine in converting wind energy into mechanical energy; power directly reflects the turbine's power generation capacity; and vibration can indicate whether the turbine has malfunctions or operational instability.

[0037] In some possible embodiments, after the sensor collects data, it can be initially processed and transmitted through a data acquisition system. The data acquisition system typically consists of a data acquisition card, signal conditioning circuitry, etc. The data acquisition card is responsible for converting the analog signals output by the sensor into digital signals for computer processing; the signal conditioning circuitry amplifies and filters the signals output by the sensor to improve signal quality and reliability. Subsequently, the acquired data is transmitted to a host computer or data center for further analysis and processing via network communication technologies such as Ethernet or wireless communication technologies (such as ZigBee, LoRa, etc.).

[0038] S102, with the optimization objectives being to maximize the total power generation efficiency of the target wind turbine group and minimize the sum of the cumulative damage of all target wind turbines, a dual-objective fitness function is constructed.

[0039] Specifically, total power generation efficiency (TGE) refers to the overall efficiency of a target wind turbine group in converting wind energy into electrical energy within a certain period. It is an important indicator for measuring the energy conversion capability of the entire wind turbine group. Improving TGE means that more electricity can be generated under the same wind energy resource conditions, thereby improving the economic benefits of the wind farm. Cumulative damage is an indicator used to assess the degree of structural damage accumulated in wind turbines due to various factors (such as wind load and vibration) during long-term operation. During operation, key components such as blades and towers of wind turbines are subjected to alternating loads, and over time, these components gradually develop fatigue damage. Cumulative damage comprehensively considers these factors and reflects the health status and service life of the wind turbines. When the sum of cumulative damage is minimized, it means that the structural damage of the entire wind turbine group is minimized, which is beneficial for extending the service life of the units and reducing maintenance costs.

[0040] Here, the bi-objective fitness function is a function that simultaneously considers the sum of the total power generation efficiency and the cumulative damage, and can be expressed as:

[0041]

[0042] Where x = [x1, x2, ..., x N ] T Represented as a set of pitch angle and yaw angle commands, x i This represents the pitch angle and yaw angle command corresponding to the i-th target wind turbine in the target wind turbine group, as shown in the pitch angle and yaw angle command set. i =(β) i ,ψ i ), β i Let ψ be the pitch angle command corresponding to the i-th target wind turbine. i Represented as the yaw angle command corresponding to the i-th target wind turbine; ω1 and ω2 are weighting coefficients, ω1+ω2=1, 0<ω1<1, 0<ω2<1; η total (x) represents the total power generation efficiency of the target wind turbine group; D i (x) represents the cumulative damage degree corresponding to the i-th target wind turbine.

[0043] Understandably, when constructing a bi-objective fitness function, each objective can be assigned a corresponding weight to reflect its importance to the overall optimization objective. The weights can be adjusted based on actual needs and experience. For example, if more emphasis is placed on power generation efficiency, the value of ω1 can be appropriately increased; if more attention is paid to the health and lifespan of the generating units, the value of ω2 can be increased. Alternatively, when market electricity prices are high, the value of ω1 can be appropriately increased; when market electricity prices are low, the value of ω2 can be increased.

[0044] S103, based on the dual-objective fitness function, the environmental data, the operating parameters of each target wind turbine, and the preset genetic algorithm, the optimal pitch angle and yaw angle instruction set of the target wind turbine group are solved.

[0045] Understandably, the optimal pitch angle and yaw angle command set includes optimal pitch angle and yaw angle commands corresponding to each target wind turbine. Here, the pitch angle refers to the angle between the wind turbine blade and the plane of rotation. Adjusting the pitch angle changes the wind force acting on the blade, thereby controlling the wind turbine's speed and power. At low wind speeds, appropriately reducing the pitch angle allows the blade to better capture wind energy, improving power generation efficiency; at high wind speeds, increasing the pitch angle limits the wind force acting on the blade, preventing turbine overload. The yaw angle refers to the angle of the wind turbine nacelle relative to the wind direction. Adjusting the yaw angle ensures the nacelle is always aligned with the wind direction, allowing the blades to receive wind energy to the maximum extent, improving aerodynamic efficiency.

[0046] Here, the genetic algorithm is defined as an optimization algorithm that simulates natural selection and genetic mechanisms. It searches for optimal solutions by simulating the evolutionary process of organisms. Its basic steps include population initialization, selection, crossover, and mutation. Natural selection is the process of survival of the fittest and elimination of the unfit in the struggle for existence. In the genetic algorithm, this is reflected in the selection of individuals based on fitness values, retaining those that are adapted to the environment (i.e., satisfy the optimization objective). The genetic mechanism encompasses phenomena such as gene transmission, crossover, and mutation in organisms. The genetic algorithm generates new individuals by simulating these phenomena, increasing population diversity and thus avoiding getting trapped in local optima. Specifically, refer to... Figure 2 As shown, when solving for the optimal pitch angle and yaw angle command set of the target wind turbine group based on the dual-objective fitness function, environmental data, operating parameters of each target wind turbine, and a preset genetic algorithm, the following steps S201 to S203 may be included:

[0047] S201, initializes a population containing multiple sets of pitch angle and yaw angle command sets.

[0048] Here, each set of pitch angle and yaw angle commands includes the pitch angle and yaw angle commands for each target wind turbine in the target wind turbine group. Specifically, a random generation method can be used when initializing the population. For example, assuming there are n wind turbines in the target wind turbine group, the pitch angle of each wind turbine can be set within a certain range (e.g., [0°, 90°]), and the yaw angle can be set within the range of [-180°, 180°]. A pitch angle and yaw angle are randomly generated for each wind turbine within the corresponding range. Combining the pitch angles and yaw angles of all wind turbines forms a set of pitch angle and yaw angle commands. Repeating this process m times (where m is the population size, which can be set according to the actual situation, e.g., m = 50) yields an initial population containing m sets of pitch angle and yaw angle commands.

[0049] S202, for each set of pitch angle and yaw angle command sets in the population, using a physical information neural network model, combined with the environmental data, the operating parameters of each target wind turbine, and the pitch angle and yaw angle commands corresponding to the pitch angle and yaw angle command sets, to determine the cumulative damage degree corresponding to each target wind turbine; and using a pre-trained power generation efficiency prediction model, combined with the environmental data, the operating parameters of each target wind turbine, and the pitch angle and yaw angle commands corresponding to the pitch angle and yaw angle command sets, to predict the total power generation efficiency of the target wind turbine group; and based on the cumulative damage degree corresponding to each target wind turbine and the total power generation efficiency of the target wind turbine group, using the bi-objective fitness function to calculate the fitness value corresponding to the pitch angle and yaw angle command sets.

[0050] Understandably, the Physical Information Neural Network (PIN) model combines physical knowledge with neural networks, utilizing physical laws to guide the learning and prediction of the neural network, thereby improving the model's accuracy and generalization ability. When determining the cumulative damage degree of each target wind turbine, the model inputs environmental data (such as wind speed, wind direction, and temperature), operating parameters of each target wind turbine (such as speed, power, and vibration), and pitch and yaw angle commands. For example, wind speed and wind direction affect the wind load on the wind turbine blades, while pitch and yaw angles change the stress state of the blades. These factors combined can lead to fatigue damage to components such as blades and towers. By learning from a large amount of experimental data and physical laws, the PIN model establishes a mapping relationship between input parameters and cumulative damage degree, thus accurately predicting the cumulative damage degree of each target wind turbine under different operating conditions.

[0051] Specifically, refer to Figure 3 As shown, the calculation of the cumulative damage degree corresponding to each target wind turbine may include the following steps S301 to S302:

[0052] S301, using the physical information neural network model, combined with the environmental data, the operating parameters of each target wind turbine, and the pitch angle and yaw angle corresponding to the population, the aerodynamic load spectrum corresponding to each target wind turbine is determined.

[0053] Here, the aerodynamic load spectrum is a curve or data set describing the magnitude and variation of aerodynamic loads experienced by a wind turbine at different times. It reflects the dynamic characteristics of the aerodynamic forces experienced by the wind turbine during operation and can be used to assess fatigue damage and reliability of the turbine. Using the aforementioned physical information neural network model, combined with environmental data, the operating parameters of each target wind turbine, and the pitch and yaw angles corresponding to the population, the aerodynamic load spectrum corresponding to each target wind turbine can be obtained.

[0054] S302, For each target wind turbine, calculate the cumulative damage degree of the target wind turbine based on the aerodynamic load spectrum and material SN curve of the target wind turbine.

[0055] Specifically, the material SN curve describes the relationship between stress amplitude (S) and fatigue life (N) of a material under alternating loads. It reflects the number of cyclic loading cycles a material can withstand at different stress levels and is an important basis for evaluating the fatigue performance of materials. In wind turbines, various components typically bear alternating aerodynamic loads, and the material SN curve can be used to predict the fatigue damage of components under specific loads. Cumulative damage is an indicator that measures the degree of fatigue damage gradually accumulated in wind turbine components due to alternating loads during long-term operation. It comprehensively considers the magnitude of the load, the number of cycles, and the fatigue characteristics of the material. When the cumulative damage reaches a certain level, the component may fail due to fatigue.

[0056] Here, the material SN curves of each component of the target wind turbine can be obtained through experimental testing. For example, fatigue tests can be conducted on the wind turbine blade material, and material samples can be cyclically loaded under different stress amplitudes. The number of cycles at which fatigue failure occurs can be recorded, and the relationship curve between stress amplitude and fatigue life can be plotted, i.e., the material SN curve. For example, for each target wind turbine, the aerodynamic load spectrum obtained in step S301 is discretized into a series of load cycles. For each load cycle, its stress amplitude is determined, and then the fatigue life N at the corresponding stress amplitude is found according to the material SN curve. Furthermore, Miner's linear cumulative damage theory can be used to calculate the cumulative damage degree. This theory states that when a material is subjected to multiple cyclic loads with different stress amplitudes, its total cumulative damage degree D is equal to the sum of the damage degrees at each stress amplitude.

[0057] In some possible implementations, a physical information neural network model can be built using deep learning frameworks (such as TensorFlow or PyTorch). First, a large amount of wind turbine operation data is collected, including environmental data, operating parameters, and actual cumulative damage (obtainable through non-destructive testing or theoretical calculations), and the data is preprocessed, such as by normalization and standardization. Then, the neural network structure is designed, including input layers, hidden layers, and output layers. Hidden layers can be fully connected layers, convolutional layers, etc., and physical constraints (such as relevant formulas from fatigue damage theory) are introduced to guide network training. The model is trained using the collected data, and the model parameters are adjusted using optimization algorithms (such as stochastic gradient descent) to minimize the error between the model's predictions and actual values.

[0058] Specifically, the pre-trained power generation efficiency prediction model is built by learning from and analyzing a large amount of wind turbine operating data. It can predict the overall power generation efficiency of a target wind turbine group based on input environmental data, the operating parameters of each target wind turbine, and pitch and yaw angle commands. Here, the power generation efficiency of wind turbines is affected by various factors, such as wind speed, wind direction, pitch angle, and yaw angle. For example, when the wind speed is within a certain range, appropriately adjusting the pitch and yaw angles can help wind turbines better capture wind energy and improve power generation efficiency. By learning the relationship between these factors and power generation efficiency, the power generation efficiency prediction model can achieve accurate prediction of the overall power generation efficiency.

[0059] In some possible implementations, a deep learning framework can be used to build a power generation efficiency prediction model. This involves collecting wind turbine operating data covering different environmental conditions and operating modes, including environmental data, operating parameters, and actual total power generation efficiency, and performing data preprocessing. The neural network architecture is then designed, employing methods such as Multilayer Perceptron (MLP) or Long Short-Term Memory (LSTM). LSTM is suitable for processing data with time-series characteristics; if operating parameters change over time, LSTM can better capture the time dependencies in the data. The model is then trained using the collected data, employing an appropriate loss function (such as mean squared error loss) and optimization algorithms to adjust the model's parameters so that the model's predictions are as close as possible to the actual total power generation efficiency.

[0060] Furthermore, after obtaining the cumulative damage degree corresponding to each target wind turbine and the total power generation efficiency of the target wind turbine group, the fitness value of the target wind turbine group under the current pitch angle and yaw angle command set can be calculated based on the bi-objective fitness function. That is, according to the formula of the bi-objective fitness function, the previously calculated cumulative damage degree and total power generation efficiency are substituted into the formula for calculation to obtain the fitness value corresponding to the pitch angle and yaw angle command set.

[0061] S203, based on the preset genetic algorithm and the fitness values ​​corresponding to each set of pitch angle and yaw angle command sets, perform an evolutionary operation on the population, and return to step S202 based on the evolved population until the preset termination condition is met, and take the pitch angle and yaw angle command set with the smallest fitness value during the evolution process as the optimal pitch angle and yaw angle command set.

[0062] Here, refer to Figure 4 As shown, when performing evolutionary operations on a population, the following steps S401 to S405 may be included:

[0063] S401, based on the fitness values ​​corresponding to the pitch angle and yaw angle instruction sets of each group in the population, determine the optimal pitch angle and yaw angle instruction set for a single wheel.

[0064] Understandably, after calculating the fitness values ​​of each pitch and yaw angle command set in the population, the minimum value is found by iterating through and comparing all fitness values. The pitch and yaw angle command set corresponding to this minimum value is the optimal pitch and yaw angle command set for a single wheel. For example, if there are 10 sets of pitch and yaw angle command sets in the population, and their calculated fitness values ​​are [0.2, 0.3, 0.5, 0.1, 0.4, 0.6, 0.25, 0.35, 0.45, 0.55], the minimum value is 0.1, and the corresponding first set of pitch and yaw angle command sets is the optimal command set for a single wheel.

[0065] S402 employs a tournament selection strategy to select multiple parent pitch angle and yaw angle instruction sets from the population.

[0066] Specifically, tournament selection is a commonly used selection method to choose superior individuals from a population as parents for subsequent genetic operations. The basic idea is to randomly select a certain number of individuals from the population, compare their fitness values, and select the individual with the lowest fitness value (in this case, the lowest) as the parent. This process is repeated multiple times until a sufficient number of parent individuals are selected. This method can, to some extent, avoid the selection results being affected by the excessively high or low fitness values ​​of individual individuals, ensuring fairness and diversity in the selection process.

[0067] For example, assuming a population size of N = 50, and k = 10 parent pitch and yaw angle command sets need to be selected, with m = 3 individuals selected for each tournament. First, three different pitch and yaw angle command sets are randomly selected from the population, and their fitness values ​​are used for selection. For example, the fitness values ​​of the three randomly selected command sets are 0.3, 0.5, and 0.2, respectively. After comparison, the command set with a fitness value of 0.2 is selected as the first parent. Then, three more different command sets are randomly selected from the population (which may overlap with the first selection), and the above comparison and selection process is repeated until 10 parent pitch and yaw angle command sets are selected.

[0068] In some other embodiments, when selecting the parent pitch angle and yaw angle instruction sets, a roulette wheel selection method can also be selected. This method determines the probability of an individual being selected based on the proportion of the individual's fitness value to the total fitness value of the population. The smaller the fitness value of an individual, the greater the probability of it being selected. No specific limitation is made here.

[0069] S403, for each selected parent pitch angle and yaw angle instruction set, perform a simulated binary crossover operation on the parent pitch angle and yaw angle instruction set to obtain the child pitch angle and yaw angle instruction set corresponding to the parent pitch angle and yaw angle instruction set.

[0070] Here, simulated binary crossover is a crossover method used in genetic algorithms with real-number encoding. It simulates the idea of ​​single-point crossover in binary encoding, but operates in the real-number space. Through this operation, partial genes of two parent individuals can be exchanged to produce new offspring individuals, thereby increasing the diversity of the population and helping to find better solutions.

[0071] S404, for each of the said child pitch angle and yaw angle instruction sets, perform a polynomial mutation operation to obtain multiple new pitch angle and yaw angle instruction sets.

[0072] Among them, polynomial mutation is a commonly used mutation method in real-number encoded genetic algorithms. It adds a random perturbation to the individual's genes, causing the individual to perform a local search in the solution space. This helps to escape local optima, increase the diversity of the population, and improve the algorithm's global search capability.

[0073] S405, the multiple new sets of pitch angle and yaw angle instruction sets and the single-wheel optimal pitch angle and yaw angle instruction set are used as the evolved population.

[0074] Specifically, the evolved population refers to a new generation of population obtained through a series of evolutionary operations such as selection, crossover, and mutation. It includes the superior individuals from the previous generation (the optimal single-wheel pitch and yaw angle command sets) and new individuals generated through genetic operations (multiple new sets of pitch and yaw angle command sets). The evolved population will serve as the initial population for the next round of evolution, continuing the optimization search and gradually approaching the optimal solution. Assume that 10 parent pitch and yaw angle command sets were selected in step S402. After the crossover and mutation operations in steps S403 and S404, 20 new sets of pitch and yaw angle command sets were obtained. Simultaneously, the optimal single-wheel pitch and yaw angle command set was determined in step S401. These 20 new command sets and the optimal single-wheel command set are merged to form a new population containing 21 command sets. This new population is the evolved population and will be used for the next round of evolutionary operations.

[0075] Furthermore, after continuous iterations based on the evolved population, when a preset termination condition is met, the pitch angle and yaw angle instruction set with the smallest fitness value during the evolution process can be taken as the optimal pitch angle and yaw angle instruction set (the smaller the fitness value, the better the solution). Here, the preset termination condition can include at least one of the following: the number of iterations of the population with the pitch angle and yaw angle instruction sets reaches a preset maximum number of iterations, the rate of change of fitness value is less than a preset rate of change of fitness value, and the evolved population has not been updated for a preset number of consecutive iterations compared to the population before evolution. For example, the maximum number of iterations is set to 100, and iteration stops when the algorithm reaches 100 iterations; or the fitness value change threshold is set to 0.001, and the algorithm is considered to have converged and iteration stops when the fitness value change is less than 0.001 for 10 consecutive generations.

[0076] S104, the optimal pitch angle and yaw angle commands corresponding to each target wind turbine are sent to the terminal layer of the corresponding target wind turbine for execution, so as to realize the coordinated action of each target wind turbine in the target wind turbine group.

[0077] Specifically, the terminal layer of the target wind turbine refers to the underlying control system of the wind turbine. It is responsible for receiving instructions from the host computer or control center and controlling the various actuators of the wind turbine (such as the pitch angle adjustment mechanism and yaw drive mechanism) to perform corresponding actions. Sending the optimal pitch angle and yaw angle instructions to the terminal layer for execution can be achieved through network communication technology. For example, industrial Ethernet or fieldbus technology (such as Profibus, Modbus, etc.) can be used to transmit instructions from the control center to the terminal controller of the wind turbine. After receiving the instructions, the terminal controller parses and processes them, and then controls the pitch angle adjustment mechanism and yaw drive mechanism to adjust the pitch angle and yaw angle according to the instructions. The pitch angle adjustment mechanism typically uses a hydraulic system or an electric pitch control system. The hydraulic system changes the pitch angle by extending and retracting a hydraulic cylinder, while the electric pitch control system adjusts the pitch angle by using a motor-driven gear transmission device. The yaw drive mechanism generally consists of a yaw motor, a reducer, and a yaw bearing. The yaw motor drives the yaw bearing to rotate through the reducer, thereby enabling the nacelle to yaw.

[0078] In this way, by having each target wind turbine unit coordinate its actions according to the optimal instructions, the entire wind turbine group can maintain high aerodynamic performance and power generation efficiency under different environmental conditions, while reducing the cumulative damage to the units and extending their service life.

[0079] In some possible embodiments, the terminal controller can be a programmable logic controller (PLC) or an embedded control system. PLCs offer advantages such as high reliability and ease of programming, enabling precise control of wind turbine actuators. Embedded control systems, on the other hand, can be customized to meet specific needs, offering greater flexibility and performance. Regarding communication protocols, the appropriate protocol must be selected based on the communication technology employed, such as Modbus TCP for industrial Ethernet communication and Modbus RTU for serial communication. Furthermore, to ensure accurate transmission and execution of instructions, communication testing and fault diagnosis are necessary to promptly identify and address any problems that arise during communication.

[0080] In some possible embodiments, refer to Figure 5 As shown, in order to further improve the safety and reliability of wind turbine operation, accurately grasp the actual status of wind turbines during operation, and promptly identify potential safety hazards, thereby rationally arranging maintenance work, reducing maintenance costs, and extending the service life of the turbines, this disclosure also proposes a wind turbine operation status monitoring and maintenance early warning method. That is, after issuing the optimal pitch angle and yaw angle commands corresponding to each target wind turbine to the terminal layer of the corresponding target wind turbine for execution, the following steps S501 to S503 may also be included:

[0081] S501, receive the latest operational sensing data about the target wind turbine group collected by sensors pre-installed in the target wind turbine group, and adjust the model parameters of the digital twin model based on the latest operational sensing data, so as to realize that the digital twin model reflects the real operating status of the target wind turbine group in real time.

[0082] Here, a digital twin model is a digital model built in virtual space based on physical models, sensor updates, historical data, etc., that completely corresponds to the actual physical entity. It can simulate and reflect the operating status, performance changes, and potential faults of an actual wind turbine group in real time, providing strong support for unit operation monitoring, optimized control, and maintenance decisions. Model parameters refer to the various variables and coefficients used in the digital twin model to define the model's behavior and characteristics, such as the unit's mass parameters, elastic modulus, and damping coefficient. These parameters affect the model's simulation effect on the actual unit's operating state.

[0083] Specifically, various sensors (such as wind speed sensors, acceleration sensors, and temperature sensors) installed at key locations in the target wind turbine cluster, such as the blade roots, main shaft, gearbox, and generator, will continuously collect data at a set sampling frequency and transmit the data to the data processing center via wired or wireless communication. Upon receiving the latest operational sensor data, the data processing center compares and analyzes it against pre-set parameters in the digital twin model. For example, if there is a discrepancy between the blade vibration data collected by the sensors and the vibration data simulated in the model based on initial parameters, it indicates that the blade dynamic parameters of the model may need adjustment. In this case, the parameters of the digital twin model can be adjusted in real time to accurately reflect the actual operating status of the target wind turbine cluster.

[0084] S502, using the material SN curve and the latest blade element load spectrum of each target wind turbine in the target wind turbine group output by the digital twin model, the cumulative damage degree of each target wind turbine is calculated in real time.

[0085] Here, the digital twin model can simulate the load changes of each blade element on each target wind turbine blade under different operating conditions based on the real-time collected operational sensor data, thereby obtaining the latest blade element load spectrum, and further calculating the cumulative damage degree corresponding to each target wind turbine (see step S302 above for details).

[0086] S503, when the cumulative damage of any target wind turbine in the target wind turbine group is greater than the preset warning threshold, a maintenance warning for the target wind turbine is triggered.

[0087] Understandably, the preset warning threshold is a critical value for cumulative damage pre-set based on factors such as the wind turbine's design requirements, material properties, operating experience, and safety factor. When the cumulative damage of the unit exceeds this threshold, it means that the unit's fatigue damage has reached a certain level, which may pose a safety hazard and requires maintenance, inspection, or repair.

[0088] Specifically, in the data processing system, a reasonable cumulative damage warning threshold is pre-set based on factors such as the wind turbine's design life, material properties, and actual operating experience. For example, it might be set to 0.8 (assuming the cumulative damage range is 0-1). After calculating the cumulative damage for each target wind turbine in real time, the system automatically compares the cumulative damage of each turbine with the preset warning threshold. If the cumulative damage of a target wind turbine exceeds 0.8, the maintenance warning mechanism is immediately triggered. Warnings can be implemented through various means, such as displaying a warning window on the monitoring center's screen showing the turbine's number, location, and current cumulative damage; simultaneously, sending warning text messages or emails to maintenance personnel's mobile phones or computer terminals to remind them to promptly inspect and maintain the turbine. Upon receiving the warning information, maintenance personnel will arrange appropriate maintenance work based on the specific situation, such as conducting a detailed inspection of the turbine and replacing damaged components, to ensure the safe operation of the turbine.

[0089] In some other embodiments of this disclosure, a forward-looking control method is proposed. This forward-looking control is based on wind speed data for the next 10 seconds obtained from LiDAR sensors at the wind farm site and the prediction of remaining fatigue life using a digital twin-deep learning model. A multi-objective optimization control method is employed to collaboratively pre-adjust the wind turbine group. The LiDAR sensors can measure wind speed and direction information at different locations in the wind farm in real time and accurately. By obtaining wind speed data for the next 10 seconds in advance, the wind conditions the wind turbines will face can be predicted. The digital twin-deep learning model combines the real-time simulation capabilities of digital twin technology with the powerful predictive capabilities of deep learning models. It can accurately predict the remaining fatigue life of the wind turbines over a future period based on historical operating data and real-time status information. Based on this predictive information, a multi-objective optimization control method is used to comprehensively consider both power generation efficiency and remaining turbine life to formulate the optimal control strategy. By collaboratively pre-adjusting the wind turbine group, such as adjusting the blade pitch angle and yaw angle in advance, the wind turbines can operate in optimal condition under future wind conditions, thereby achieving the dual objectives of increasing power generation and extending remaining life. This allows for full utilization of dynamic information from the wind farm, enabling proactive decision-making and improving the operational performance and reliability of wind turbine units.

[0090] In some possible embodiments, in order to improve the intelligence level of wind turbine operation and management, achieve efficient data processing, accurate prediction and optimized control, thereby improving power generation efficiency, extending the service life of wind turbines and reducing operation and maintenance costs, a cloud-edge-device collaborative layered architecture can be adopted, which may include a cloud layer, an edge layer and a terminal layer.

[0091] Specifically, the cloud layer, acting as the data hub and decision-making brain of the entire architecture, can undertake several key tasks. In terms of data management, it is responsible for the comprehensive storage and orderly management of wind turbine operating data and sensor data. This data covers operating parameters and environmental parameters of various components of the wind turbine, ensuring secure storage, rapid retrieval, and convenient access through an efficient database management system. Based on digital twin technology, the cloud layer can achieve a visual display of the wind turbine. By constructing a virtual model that corresponds one-to-one with the actual wind turbine, real-time operating data is mapped onto the model. Maintenance personnel can intuitively observe the operating status of the wind turbine in a visual interface, including blade rotation and unit vibration, and promptly identify potential problems.

[0092] Meanwhile, the cloud layer also possesses algorithm upgrade capabilities. It can promptly integrate these advanced algorithms into the system, upgrading and optimizing existing control and prediction models to improve system performance and accuracy. Furthermore, the cloud layer needs to interface with the centralized control system to achieve information exchange with the entire wind farm's centralized control center. Through this interface, the cloud layer can upload wind turbine operating data and status information in real time, providing decision-making support for the centralized control system; simultaneously, it can receive scheduling commands from the centralized control system, enabling remote control and coordinated management of the wind turbine units.

[0093] Specifically, the edge layer, situated between the cloud layer and the terminal layer, plays a role in data preprocessing and preliminary decision-making and control. It first receives monitoring data from the terminal layer, which may contain noise and interference. At this point, the edge layer performs signal filtering and data cleaning to remove invalid data and noise interference, improving data quality and accuracy. Furthermore, based on a lightweight physical information neural network model, the edge layer combines material SN curves to predict the remaining fatigue life of the wind turbine in real time. The physical information neural network model combines physical knowledge with data-driven methods, reducing computational load and complexity while maintaining prediction accuracy, making it suitable for deployment on edge devices. By analyzing wind turbine operating data and material performance parameters in real time, the edge layer can accurately predict the remaining fatigue life of various wind turbine components, providing early warnings of high-risk areas and timely maintenance recommendations for maintenance personnel. The edge layer is also responsible for uploading and receiving cloud commands and distributing control commands to the terminal layer. It can accurately control the wind turbine on the terminal side based on cloud decisions and local real-time data. With the goal of balancing improved power generation efficiency and extended wind turbine remaining life, the edge layer performs group collaborative optimization of the wind turbine's pitch and yaw control. By considering the mutual influence between wind turbines and the overall operating efficiency of the wind farm, the edge layer can formulate optimal control strategies to achieve coordinated operation of the wind turbine group. Furthermore, the edge layer also has fault and hazard warning functions; when abnormal conditions or potential hazards are detected in the wind turbines, it can promptly issue alarms to notify maintenance personnel for handling.

[0094] Specifically, the terminal layer is the data acquisition and command execution end of the entire architecture. It mainly consists of wind turbine sensors distributed across various key parts of the wind turbine to monitor environmental and operational data. Environmental data includes wind speed, wind direction, temperature, and humidity; wind turbine operational data covers parameters such as blade speed, unit vibration, and generator power, directly reflecting the wind turbine's operating status. The terminal layer is responsible for uploading the data collected by these sensors to the edge layer in real time, providing a foundation for upper-layer data processing and decision-making. Simultaneously, the terminal layer can also execute control commands from the edge layer. After the edge layer generates a control strategy based on optimization algorithms and real-time data, it sends the corresponding control commands to the terminal layer. The terminal layer then controls the wind turbine's actuators, such as the pitch and yaw systems, to achieve precise control of the wind turbine, ensuring it operates in its optimal state.

[0095] The real-time optimization method, apparatus, medium, and equipment for the aerodynamic performance of wind turbine units provided in this disclosure achieves real-time optimization of the aerodynamic performance of wind turbine units by constructing a dual-objective fitness function and combining it with a genetic algorithm to solve for the optimal instruction set, thereby uniformly adjusting the pitch angle and yaw angle of each unit. This effectively improves the overall power generation efficiency of the wind turbine group, while reducing the cumulative damage of the units and extending the service life of the equipment.

[0096] Those skilled in the art will understand that, in the above-described method of the specific implementation, the order in which each step is written does not imply a strict execution order and does not constitute any limitation on the implementation process. The specific execution order of each step should be determined by its function and possible internal logic.

[0097] Based on the same inventive concept, this disclosure also provides a wind turbine aerodynamic performance real-time optimization device corresponding to the wind turbine aerodynamic performance real-time optimization method. Since the principle of the device in this disclosure for solving the problem is similar to the wind turbine aerodynamic performance real-time optimization method described above, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be described again.

[0098] Reference Figure 6 The diagram shown is a schematic of a real-time aerodynamic performance optimization device 600 for wind turbines provided in an embodiment of this disclosure. The device includes:

[0099] The data acquisition module 601 is used to acquire sensing data about the target wind turbine group based on sensors pre-installed in the target wind turbine group; wherein, the sensing data includes environmental data and operating parameters of each target wind turbine;

[0100] The function construction module 602 is used to construct a dual-objective fitness function with the optimization objectives of maximizing the total power generation efficiency of the target wind turbine group and minimizing the sum of the cumulative damage of all target wind turbines.

[0101] The instruction solving module 603 is used to solve the optimal pitch angle and yaw angle instruction set of the target wind turbine group based on the dual-objective fitness function, the environmental data, the operating parameters of each target wind turbine, and a preset genetic algorithm; wherein, the optimal pitch angle and yaw angle instruction set includes the optimal pitch angle and yaw angle instructions corresponding to each target wind turbine.

[0102] The instruction issuing module 604 is used to issue the optimal pitch angle and yaw angle instructions corresponding to each target wind turbine to the terminal layer of the corresponding target wind turbine for execution, so as to realize the coordinated action of each target wind turbine in the target wind turbine group.

[0103] In some possible embodiments, the bi-objective fitness function is expressed as:

[0104]

[0105] Where x = [x1, x2, ..., x N ] T Represented as a set of pitch angle and yaw angle commands, x iThis represents the pitch angle and yaw angle command corresponding to the i-th target wind turbine in the target wind turbine group, as shown in the pitch angle and yaw angle command set. i =(β) i ,ψ i ), β i Let ψ be the pitch angle command corresponding to the i-th target wind turbine. i Represented as the yaw angle command corresponding to the i-th target wind turbine; ω1 and ω2 are weighting coefficients, ω1+ω2=1, 0<ω1<1, 0<ω2<1; η total (x) represents the total power generation efficiency of the target wind turbine group; D i (x) represents the cumulative damage degree corresponding to the i-th target wind turbine.

[0106] In some possible embodiments, the instruction solving module 603 is specifically used for:

[0107] Step 1: Initialize a population containing multiple sets of pitch angle and yaw angle command sets; wherein each set of pitch angle and yaw angle command sets includes the pitch angle and yaw angle command for each target wind turbine in the target wind turbine group;

[0108] Step 2: For each set of pitch angle and yaw angle command sets in the population, using a physical information neural network model, combined with the environmental data, the operating parameters of each target wind turbine, and the pitch angle and yaw angle commands corresponding to the set of pitch angle and yaw angle command sets, determine the cumulative damage degree corresponding to each target wind turbine; and using a pre-trained power generation efficiency prediction model, combined with the environmental data, the operating parameters of each target wind turbine, and the pitch angle and yaw angle commands corresponding to the set of pitch angle and yaw angle command sets, predict the total power generation efficiency of the target wind turbine group; and based on the cumulative damage degree corresponding to each target wind turbine and the total power generation efficiency of the target wind turbine group, use the bi-objective fitness function to calculate the fitness value corresponding to the set of pitch angle and yaw angle command sets.

[0109] Step 3: Based on the preset genetic algorithm and the fitness values ​​corresponding to each set of pitch angle and yaw angle command sets, perform evolutionary operations on the population, and return to Step 2 based on the evolved population until the preset termination condition is met. The set of pitch angle and yaw angle command sets with the smallest fitness value during the evolution process is taken as the optimal set of pitch angle and yaw angle command sets.

[0110] In some possible embodiments, the instruction solving module 603 is further configured to:

[0111] Using the physical information neural network model, combined with the environmental data, the operating parameters of each target wind turbine, and the pitch angle and yaw angle of the population, the aerodynamic load spectrum corresponding to each target wind turbine is determined.

[0112] For each target wind turbine, the cumulative damage degree of the target wind turbine is calculated based on the aerodynamic load spectrum and the material SN curve.

[0113] In some possible embodiments, the instruction solving module 603 is further configured to:

[0114] Based on the fitness values ​​corresponding to each set of pitch angle and yaw angle commands in the population, the optimal set of pitch angle and yaw angle commands for a single wheel is determined.

[0115] A tournament selection strategy is used to select multiple sets of parent pitch angle and yaw angle commands from the population;

[0116] For each selected parent pitch angle and yaw angle instruction set, a simulated binary crossover operation is performed on the parent pitch angle and yaw angle instruction set to obtain the child pitch angle and yaw angle instruction set corresponding to the parent pitch angle and yaw angle instruction set.

[0117] For each of the child pitch angle and yaw angle command sets, a polynomial mutation operation is performed to obtain multiple new pitch angle and yaw angle command sets.

[0118] The multiple new sets of pitch angle and yaw angle command sets and the single-wheel optimal pitch angle and yaw angle command set are used as the evolved population.

[0119] In some possible embodiments, the preset termination condition includes at least one of the following: the number of iterations of the population of pitch angle and yaw angle command sets reaches a preset maximum number of iterations, the rate of change of fitness values ​​is less than a preset rate of change of fitness values, and the evolved population has no updates for a preset number of consecutive times compared to the population before evolution.

[0120] In some possible embodiments, the instruction issuing module 604 is further configured to:

[0121] The system receives the latest operational sensing data about the target wind turbine group from sensors pre-installed in the target wind turbine group, and adjusts the model parameters of the digital twin model based on the latest operational sensing data, so that the digital twin model can reflect the real operational status of the target wind turbine group in real time.

[0122] Using the material SN curve and the latest blade element load spectrum of each target wind turbine in the target wind turbine group output by the digital twin model, the cumulative damage degree of each target wind turbine is calculated in real time.

[0123] When the cumulative damage of any target wind turbine in the target wind turbine group exceeds a preset warning threshold, a maintenance warning for the target wind turbine is triggered.

[0124] Based on the same technical concept, this disclosure also provides a computer device. (See also...) Figure 7 The diagram shows the structure of a computer device 700 provided in this embodiment of the present disclosure, including a processor 701, a memory 702, and a bus 703. The memory 702 stores execution instructions and includes a main memory 7021 and an external memory 7022. The main memory 7021, also called internal memory, is used to temporarily store computational data in the processor 701, as well as data exchanged with external memory 7022 such as a hard disk. The processor 701 exchanges data with the external memory 7022 through the main memory 7021.

[0125] In this embodiment, the memory 702 is specifically used to store application code that executes the solution of this application, and its execution is controlled by the processor 701. That is, when the computer device 700 is running, the processor 701 communicates with the memory 702 through the bus 703, so that the processor 701 executes the application code stored in the memory 702, and then executes the method described in any of the foregoing embodiments.

[0126] The memory 702 may be, but is not limited to, random access memory (RAM), read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), etc.

[0127] Processor 701 may be an integrated circuit chip with signal processing capabilities. The aforementioned processor can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this invention. The general-purpose processor can be a microprocessor or any conventional processor.

[0128] It is understood that the structures illustrated in the embodiments of this application do not constitute a specific limitation on the computer device 700. In other embodiments of this application, the computer device 700 may include more or fewer components than illustrated, or combine some components, or split some components, or have different component arrangements. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0129] This disclosure also provides a computer-readable storage medium storing a computer program. When executed by a processor, the computer program performs the steps of the real-time aerodynamic performance optimization method for wind turbines described in the above-described method embodiments. The storage medium can be either volatile or non-volatile computer-readable storage.

[0130] This disclosure also provides a computer program product carrying program code. The program code includes instructions that can be used to execute the steps of the real-time optimization method for the aerodynamic performance of wind turbines described in the above method embodiments. For details, please refer to the above method embodiments, which will not be repeated here.

[0131] The aforementioned computer program product can be implemented through hardware, software, or a combination thereof. In one optional embodiment, the computer program product is specifically embodied in a computer storage medium; in another optional embodiment, the computer program product is specifically embodied in a software product, such as a software development kit (SDK), etc.

[0132] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and devices described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. In the several embodiments provided in this disclosure, it should be understood that the disclosed systems and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division; in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Another point is that the displayed or discussed mutual coupling or direct coupling or communication connection may be through some communication interfaces; the indirect coupling or communication connection of devices or units may be electrical, mechanical, or other forms.

[0133] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0134] In addition, the functional units in the various embodiments of this disclosure can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0135] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a processor-executable, non-volatile, computer-readable storage medium. Based on this understanding, the technical solution of this disclosure, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this disclosure. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0136] Finally, it should be noted that the above-described embodiments are merely specific implementations of this disclosure, used to illustrate the technical solutions of this disclosure, and not to limit it. The protection scope of this disclosure is not limited thereto. Although this disclosure has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features, within the scope of the technology disclosed in this disclosure. Such modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this disclosure, and should all be covered within the protection scope of this disclosure. Therefore, the protection scope of this disclosure should be determined by the protection scope of the claims.

Claims

1. A method for real-time optimization of the aerodynamic performance of a wind turbine generator set, characterized in that, include: Sensor data about the target wind turbine group is collected based on sensors pre-installed in the target wind turbine group; wherein, the sensor data includes environmental data and operating parameters of each target wind turbine group; The optimization objectives are to maximize the total power generation efficiency of the target wind turbine group and minimize the sum of the cumulative damage of all target wind turbines, and a dual-objective fitness function is constructed. Based on the bi-objective fitness function, the environmental data, the operating parameters of each target wind turbine, and the preset genetic algorithm, the optimal pitch angle and yaw angle command set of the target wind turbine group is solved; wherein, the optimal pitch angle and yaw angle command set includes the optimal pitch angle and yaw angle command corresponding to each target wind turbine. The optimal pitch angle and yaw angle commands corresponding to each target wind turbine are sent to the terminal layer of the corresponding target wind turbine for execution, so as to realize the coordinated operation of each target wind turbine in the target wind turbine group.

2. The method according to claim 1, characterized in that, The bi-objective fitness function is expressed as follows: Where x = [x1, x2, ..., x N ] T Represented as a set of pitch angle and yaw angle commands, x i This represents the pitch angle and yaw angle command corresponding to the i-th target wind turbine in the target wind turbine group, as shown in the pitch angle and yaw angle command set. i =(β) i ,ψ i ), β i Let ψ be the pitch angle command corresponding to the i-th target wind turbine. i Represented as the yaw angle command corresponding to the i-th target wind turbine; ω1 and ω2 are weighting coefficients, ω1+ω2=1, 0<ω1<1, 0<ω2<1; η total (x) represents the total power generation efficiency of the target wind turbine group; D i (x) represents the cumulative damage degree corresponding to the i-th target wind turbine.

3. The method according to claim 2, characterized in that, The process of solving for the optimal pitch angle and yaw angle command set of the target wind turbine group based on the dual-objective fitness function, the environmental data, the operating parameters of each target wind turbine, and a preset genetic algorithm includes: Step 1: Initialize a population containing multiple sets of pitch angle and yaw angle command sets; wherein each set of pitch angle and yaw angle command sets includes the pitch angle and yaw angle command for each target wind turbine in the target wind turbine group; Step 2: For each set of pitch angle and yaw angle command sets in the population, using a physical information neural network model, combined with the environmental data, the operating parameters of each target wind turbine, and the pitch angle and yaw angle commands corresponding to the set of pitch angle and yaw angle command sets, determine the cumulative damage degree corresponding to each target wind turbine; and using a pre-trained power generation efficiency prediction model, combined with the environmental data, the operating parameters of each target wind turbine, and the pitch angle and yaw angle commands corresponding to the set of pitch angle and yaw angle command sets, predict the total power generation efficiency of the target wind turbine group; and based on the cumulative damage degree corresponding to each target wind turbine and the total power generation efficiency of the target wind turbine group, use the bi-objective fitness function to calculate the fitness value corresponding to the set of pitch angle and yaw angle command sets. Step 3: Based on the preset genetic algorithm and the fitness values ​​corresponding to each set of pitch angle and yaw angle command sets, perform evolutionary operations on the population, and return to Step 2 based on the evolved population until the preset termination condition is met. The set of pitch angle and yaw angle command sets with the smallest fitness value during the evolution process is taken as the optimal set of pitch angle and yaw angle command sets.

4. The method according to claim 3, characterized in that, The method of using a physical information neural network model, combined with environmental data, operating parameters of each target wind turbine, and pitch and yaw angle commands corresponding to the pitch and yaw angle command set, to determine the cumulative damage degree of each target wind turbine includes: Using the physical information neural network model, combined with the environmental data, the operating parameters of each target wind turbine, and the pitch angle and yaw angle of the population, the aerodynamic load spectrum corresponding to each target wind turbine is determined. For each target wind turbine, the cumulative damage degree of the target wind turbine is calculated based on the aerodynamic load spectrum and the material SN curve.

5. The method according to claim 3, characterized in that, The evolutionary operation performed on the population based on a preset genetic algorithm and the fitness values ​​corresponding to each set of pitch angle and yaw angle command sets includes: Based on the fitness values ​​corresponding to each set of pitch angle and yaw angle commands in the population, the optimal set of pitch angle and yaw angle commands for a single wheel is determined. A tournament selection strategy is used to select multiple sets of parent pitch angle and yaw angle commands from the population; For each selected parent pitch angle and yaw angle instruction set, a simulated binary crossover operation is performed on the parent pitch angle and yaw angle instruction set to obtain the child pitch angle and yaw angle instruction set corresponding to the parent pitch angle and yaw angle instruction set. For each of the child pitch angle and yaw angle command sets, a polynomial mutation operation is performed to obtain multiple new pitch angle and yaw angle command sets. The multiple new sets of pitch angle and yaw angle command sets and the single-wheel optimal pitch angle and yaw angle command set are used as the evolved population.

6. The method according to claim 5, characterized in that, The preset termination conditions include at least one of the following: the number of iterations of the population of pitch angle and yaw angle command sets reaches a preset maximum number of iterations, the rate of change of fitness values ​​is less than a preset rate of change of fitness values, and the evolved population has no updates for a preset number of consecutive iterations compared to the population before evolution.

7. The method according to claim 1, characterized in that, After the optimal pitch angle and yaw angle commands corresponding to each target wind turbine are sent to the terminal layer of the corresponding target wind turbine for execution, the process further includes: The system receives the latest operational sensing data about the target wind turbine group from sensors pre-installed in the target wind turbine group, and adjusts the model parameters of the digital twin model based on the latest operational sensing data, so that the digital twin model can reflect the real operational status of the target wind turbine group in real time. Using the material SN curve and the latest blade element load spectrum of each target wind turbine in the target wind turbine group output by the digital twin model, the cumulative damage degree of each target wind turbine is calculated in real time. When the cumulative damage of any target wind turbine in the target wind turbine group exceeds a preset warning threshold, a maintenance warning for the target wind turbine is triggered.

8. A real-time aerodynamic performance optimization device for wind turbine generators, characterized in that, include: The data acquisition module is used to acquire sensing data about the target wind turbine group based on sensors pre-installed in the target wind turbine group; wherein, the sensing data includes environmental data and operating parameters of each target wind turbine; The function construction module is used to construct a dual-objective fitness function with the optimization objectives of maximizing the total power generation efficiency of the target wind turbine group and minimizing the sum of the cumulative damage of all target wind turbines. The instruction solving module is used to solve for the optimal pitch angle and yaw angle instruction set of the target wind turbine group based on the dual-objective fitness function, the environmental data, the operating parameters of each target wind turbine, and a preset genetic algorithm; wherein, the optimal pitch angle and yaw angle instruction set includes the optimal pitch angle and yaw angle instructions corresponding to each target wind turbine. The instruction issuing module is used to issue the optimal pitch angle and yaw angle instructions corresponding to each target wind turbine to the terminal layer of the corresponding target wind turbine for execution, so as to realize the coordinated action of each target wind turbine in the target wind turbine group.

9. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 7.

10. A computer device, comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1 to 7.

Images