Wake flow evaluation method, device and equipment for floating wind power plant and storage medium

Abstract

The invention relates to the technical field of wind power plants, in particular to a wake flow evaluation method and device for a floating wind power plant, equipment and a storage medium. According to the method, high precision of numerical simulation and high efficiency of data driving are combined, namely, CFD is firstly adopted for simulation, and then a symbol regression model based on a particle swarm optimization algorithm is trained based on data obtained through simulation, so that a wake flow velocity loss expression is constructed, and the influence of wake flow interference existing in the floating wind power plant is simulated. And meanwhile, compared with the traditional method that fan integrated numerical calculation is carried out by using CFD, the calculation time consumption and the cost are very high, the method carries out multi-equation comparison through a particle swarm optimization algorithm, an optimal symbol regression equation is optimized, and rapid prediction of a key region of interest in engineering is realized. Besides, for the wake flow area, a smooth step function is adopted to replace an indicator function, so that the problem that the wake flow expansion rate contains jump discontinuity, and sharp change of the wake flow width can be caused can be avoided.

Description

Technical Field

[0001] This invention relates to the field of wind farm technology, and specifically to a method, apparatus, equipment, and storage medium for assessing the wake of a floating wind farm. Background Technology

[0002] Wind power generation converts wind kinetic energy into electrical energy through technological means, thereby achieving the effective utilization of wind energy. As a clean energy source, it has advantages such as environmental friendliness, renewability, and low operation and maintenance costs. Currently, wind resource acquisition is gradually shifting from onshore to offshore, and offshore development is also gradually moving towards the deep sea. However, the complex offshore environment places high demands on the stability of the power system. The volatility and intermittency of wind energy not only affect power quality but also relate to the safe operation and economic dispatch of the power grid. As a core component of addressing these challenges, wind resource forecasting is crucial throughout the entire lifecycle of a wind farm: from early site selection to avoid meteorological disasters to optimizing maintenance plans and mitigating extreme weather during the operational phase, its accuracy directly determines the survivability and profitability of the wind farm.

[0003] Floating wind farms are typically constructed in a spaced array configuration, but numerical simulations indicate that wake losses from upstream turbines significantly impact the performance of downstream turbines. Therefore, research on the wake characteristics of floating wind turbines is essential for understanding the overall power generation and efficiency of wind farms.

[0004] Currently, research methods for wake structures include wind tunnel testing, field measurements, numerical simulation, and data-driven methods. Among these, wind tunnel testing is costly in terms of time, manpower, and materials from design to implementation. Furthermore, the experimental environment is limited, making it difficult to achieve satisfactory results, and the results have low reusability and cannot form a systematic understanding. In field measurements, the coverage of marine meteorological stations and wind measurement towers is sparse, making data acquisition difficult and resulting in poor spatiotemporal continuity of measured data.

[0005] Computational Fluid Dynamics (CFD) methods in numerical simulations can acquire comprehensive flow field data and offer environmental controllability, but their computational cost is high and they are difficult to apply to the dynamic motion of floating wind turbines under complex marine environmental loads. Meanwhile, the limitations of traditional wake models further exacerbate the challenges: in terms of physical models, classic wake models such as Jensen's cannot characterize the dynamic wake distortion caused by the motion of floating wind turbines; in terms of computational efficiency, although CFD can simulate complex flow fields, it takes several hours per single case, making it difficult to meet real-time control requirements; and in terms of extrapolation capabilities, the diverse layouts of offshore wind farms cause fixed empirical formulas to show a sharp increase in error under the new floating turbine scenario.

[0006] Data-driven methods are computationally efficient and can handle a variety of complex scenarios, but they require a large amount of data, and pure data-driven models have poor interpretability and insufficient generalization ability.

[0007] In summary, the above methods have significant limitations, mainly including the following aspects:

[0008] (1) High time cost. High-fidelity CFD models can capture flow details, but the computation time and cost are often high. Taking a simulation with a million-grid density as an example, small time steps are often needed to control the progress of numerical calculation, and a certain time period is required to reproduce the relevant phenomena. If the calculation involves a large wind farm, even with the use of supercomputer resources, a very long calculation period is still required.

[0009] (2) It is difficult to verify the effect and optimize the parameterization scheme. Due to the lack of marine observation data and the poor spatiotemporal continuity of the data, it is difficult to use measured data to evaluate the wake and further optimize the parameterization scheme.

[0010] (3) Data structure errors accumulate gradually, and the accuracy of the data source is insufficient. The performance of machine learning models is highly dependent on the quality and quantity of training data. If the accuracy of the data source is insufficient, the model may not be able to learn accurate tail characteristics or generalize well to unseen data or environmental conditions, resulting in inaccurate predictions. Summary of the Invention

[0011] In view of this, the present invention provides a method, apparatus, equipment and storage medium for wake assessment of floating wind farms, in order to solve the problems of high time cost and poor adaptability of wake research in the prior art.

[0012] In a first aspect, the present invention provides a method for evaluating the wake of a floating wind farm. The method includes: simulating the operation of the floating wind turbine using computational fluid dynamics to obtain wake flow field data; replacing the indicator function characterizing the wake region with a smoothing step function based on the wake flow field data; determining the wake velocity deficit expression using a symbolic regression model based on particle swarm optimization algorithm according to the wake flow field data adjusted for the wake region; and evaluating the wake of the floating wind farm based on the wake velocity deficit expression.

[0013] This invention combines the high precision of numerical simulation with the high efficiency of data-driven approaches. First, CFD simulation is employed. Then, based on the simulation data, a symbolic regression model based on particle swarm optimization (PSO) is trained to construct a wake velocity deficit expression, thereby simulating the impact of wake interference in floating wind farms. Furthermore, compared to the traditional method of using CFD for integrated wind turbine numerical calculations, which is time-consuming and costly, this method uses PSO to compare multiple equations and select the optimal symbolic regression equation, enabling rapid prediction of key engineering areas of concern. In addition, for the wake region, a smoothing step function is used instead of the indicator function to avoid the problem of sharp changes in wake width caused by jumps and discontinuities in the wake spread rate.

[0014] In one optional implementation, computational fluid dynamics (CFD) is used to simulate the operation of the floating wind turbine to obtain wake flow field data. This includes: selecting simulation software based on CFD; using the simulation software to model the floating wind turbine, wind farm layout, and parameters under different inflow conditions; determining the numerical solution method and boundary conditions, and refining the mesh of the wake region; correcting the model parameters obtained from the modeling, and solving for the aerodynamic performance to obtain wake flow field data.

[0015] In this invention, by selecting and adapting simulation software, various inflow conditions, layouts and parameters are modeled, the wake mesh is refined, the model parameters are corrected and the aerodynamic performance is solved. This enables the accurate acquisition of wake flow field data, providing high-quality basic data for subsequent modeling such as symbolic regression, and improving the accuracy and reliability of wake characteristic research and engineering applications.

[0016] In one optional implementation, the wake region is represented by the interval [a, b). The indicator function indicates whether any point is within the wake region; if it is within the wake region, it returns 1; otherwise, it returns 0. The smoothing step function includes the intervals [ad, a+d], [a+d, bd], [bd, b+d], and (-∞, ad) and (b+d, +∞). Within the interval [ad, a+d], the smoothing step function smoothly increases from 0 to 1; within the interval [a+d, bd], the smoothing step function returns 1; within the interval [bd, b+d], the smoothing step function smoothly decreases from 1 to 1; and within the intervals (-∞, ad) and (b+d, +∞), the smoothing step function returns 0. Here, d represents the standard deviation of the Gaussian distribution of the distance from any point in the wake region to the starting point of the wake region.

[0017] In this invention, by setting multiple intervals and specific variations of the smoothing step function, a smooth transition of the wake region boundary is achieved, avoiding the abrupt change problem of traditional indicator functions. Combined with the Gaussian distribution standard deviation d, it can better fit the actual physical characteristics of wake diffusion, providing a more reasonable and continuous region definition for subsequent wake modeling (such as symbolic regression), and improving model accuracy and physical consistency.

[0018] In one optional implementation, the wake velocity deficit expression is determined using a symbolic regression model based on particle swarm optimization (PSO) algorithm, based on the wake flow field data adjusted for the wake region. This includes: acquiring model training data based on the wake flow field data adjusted for the wake region, wherein the model training data includes floating wind turbine layout parameters, amplitude, frequency, and normalized velocity deficit; constructing the search space of the symbolic regression model and generating an initial expression as particles for PSO optimization; encoding, evaluating the fitness of the particles, and optimizing them to determine the wake velocity deficit expression.

[0019] In one alternative implementation, when evaluating the fitness of particles, the minimum wake loss at the heave ratio and yaw rate is used as the fitness value, and the minimum wake loss rate is used as the guide.

[0020] In this invention, based on the adjusted wake flow field data, by acquiring multi-dimensional training data, constructing a search space, encoding and optimizing particles, the wake velocity loss pattern can be accurately discovered, and an expression adapted to the floating wind turbine scenario can be generated. This takes into account both physical mechanisms and data fitting, providing efficient and accurate model support for wake characteristic analysis.

[0021] In one alternative implementation, the relationship between the heave ratio and yaw rate and the wake loss is expressed by the following formula:

[0022]

[0023] In the formula, U max denoted as , where z is the maximum wind speed within the impeller sweep surface, and z is the heave displacement. For the yaw angle, U ∞ The upstream wind speed is the source of the interference from the primary wind turbine, and dz / dt is the heave velocity. This is the yaw rate.

[0024] In this invention, the relationship between heave velocity ratio, yaw rate and wake loss is quantified by formula. The dynamic motion parameters (heave and yaw) of the floating wind turbine are integrated into the wake loss analysis, which can accurately characterize the impact of motion on the wake, provide dynamic physical constraints for floating wind turbine wake modeling, and help the symbolic regression model capture more realistic wake patterns.

[0025] In one optional implementation, evaluating the wake of a floating wind farm based on the wake velocity deficit expression includes: evaluating the wake loss under different blade deformation conditions based on the wake velocity deficit expression; the method further includes: quantifying the wake loss based on the power output of the wind farm and the power output of a single floating wind turbine.

[0026] In this invention, based on the evaluation of wind flow using the wake velocity loss expression, the wake loss is further evaluated using a wind farm power model and an equivalent model of a single wind turbine.

[0027] Secondly, the present invention provides a wake assessment device for a floating wind farm. The device includes: a simulation module for simulating the operation of the floating wind turbine using computational fluid dynamics to obtain wake flow field data; a function replacement module for replacing the indicator function characterizing the wake region with a smoothing step function based on the wake flow field data; an expression determination module for determining the wake velocity deficit expression using a symbolic regression model based on particle swarm optimization algorithm according to the wake flow field data adjusted for the wake region; and an assessment module for assessing the wake of the floating wind farm based on the wake velocity deficit expression.

[0028] Thirdly, the present invention provides a computer device, comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the computer instructions to perform the wake assessment method for floating wind farms described in the first aspect or any corresponding embodiment thereof.

[0029] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to execute the wake assessment method for a floating wind farm according to the first aspect or any corresponding embodiment described above.

[0030] Fifthly, the present invention provides a computer program product, including computer instructions for causing a computer to execute the wake assessment method for a floating wind farm according to the first aspect or any corresponding embodiment described above. Attached Figure Description

[0031] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0032] Figure 1This is a flowchart illustrating the wake assessment method for a floating wind farm according to an embodiment of the present invention.

[0033] Figure 2 This is a schematic diagram of the wake flow field data acquisition process according to an embodiment of the present invention;

[0034] Figure 3 This is a flowchart illustrating the particle swarm optimization algorithm according to an embodiment of the present invention;

[0035] Figure 4 This is a flowchart illustrating another method for evaluating the wake of a floating wind farm according to an embodiment of the present invention.

[0036] Figure 5 This is a structural block diagram of a wake assessment device for a floating wind farm according to an embodiment of the present invention;

[0037] Figure 6 This is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. Detailed Implementation

[0038] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0039] According to an embodiment of the present invention, a method for evaluating the wake of a floating wind farm is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0040] This embodiment provides a method for evaluating the wake of a floating wind farm, which can be used in electronic devices such as computers, mobile phones, and tablets. Figure 1 This is a flowchart of a method for evaluating the wake of a floating wind farm according to an embodiment of the present invention, as shown below. Figure 1 As shown, the process includes the following steps:

[0041] Step S101 involves simulating the operation of the floating wind turbine using computational fluid dynamics (CFD) to obtain wake flow field data. Specifically, simulation software can be used to input relevant parameters of the floating wind turbine, such as its operating parameters, to obtain the corresponding wake flow field data. Furthermore, since the wake loss of the upstream wind turbine significantly affects the performance of the downstream wind turbine, the simulation actually simulates the operation of a floating wind farm composed of multiple floating wind turbines. Simultaneously, compared to traditional CFD simulations which require smaller time steps to control the numerical calculation process and a certain time period to reproduce related phenomena, this embodiment uses CFD simulation only to provide training data for the symbolic regression model. Therefore, the time step and calculation period settings only need to meet the requirements of the training data, such as obtaining sufficient static or quasi-static flow field characteristics covering key operating conditions, thereby reducing time costs.

[0042] Step S102: Based on the wake flow field data, the indicator function characterizing the wake region is replaced with a smooth step function. Specifically, the wake flow field data includes the flow field data of the wake region, which is typically defined using an exponential function. The wake region is usually a defined physical region, such as the interval from a to b. Within this interval, the characteristics of the wake (such as velocity, pressure, etc.) differ from those outside the interval. The indicator function 1[a,b)(x) is used to explicitly define this region. For example, assuming the width of the wake in a certain direction is from a to b, the indicator function can be used to indicate whether a point x is within the wake region. If x is within [a,b), the indicator function returns 1, indicating that the point is within the wake region; otherwise, it returns 0. The smooth step function, on the other hand, is a continuous function used to replace the indicator function to avoid discontinuities. The smooth step function smoothly changes from 0 to 1 within the interval [a,b), rather than jumping abruptly from 0 to 1 like the indicator function. This avoids the problem of "sharp" changes in wake width caused by jumps and discontinuities in the wake spread rate.

[0043] Step S103: Based on the adjusted wake flow field data in the wake region, a symbolic regression model based on particle swarm optimization algorithm is used to determine the wake velocity deficit expression.

[0044] Step S104: Evaluate the wake of the floating wind farm based on the wake velocity deficit expression.

[0045] Specifically, in floating wind farms, considering the complexity and diversity of the environment in which floating wind turbines operate, they may be subject to the combined effects of wind, waves, currents, and other external loads. Furthermore, the operation of the upper turbines and the platform, mooring, and other systems will also interact. Therefore, reasonable layout and optimization are crucial for the efficient operation of both the turbines and the wind farm. For simulating single turbines and interactions between turbines, constructing accurate parameter models is fundamental for efficient and high-precision calculations. In this embodiment, a symbolic regression model using particle swarm optimization is employed to determine the wake velocity deficit expression as the wake model for evaluating the wake of the floating wind farm. The symbolic regression model determines the search space, i.e., the basic operator set of the symbolic regression model. Then, the particle swarm optimization algorithm is used to select appropriate operators and functions, and to determine the final expression.

[0046] The wake assessment method for floating wind farms provided in this invention combines the high precision of numerical simulation with the high efficiency of data-driven approaches. First, CFD simulation is performed. Then, a symbolic regression model based on particle swarm optimization (PSO) is trained using the simulation data to construct a wake velocity deficit expression, thereby simulating the impact of wake interference in floating wind farms. Compared to traditional CFD-based integrated wind turbine numerical calculations, which are time-consuming and costly, this method uses PSO to compare multiple equations and select the optimal symbolic regression equation, enabling rapid prediction of key engineering areas of concern. Furthermore, for the wake region, a smoothing step function is used instead of the indicator function to avoid the problem of sharp changes in wake width caused by jumps and discontinuities in the wake spread rate.

[0047] This embodiment provides a wake assessment method for floating wind farms, the process of which includes the following steps:

[0048] Step S201: The operation of the floating fan is simulated using computational fluid dynamics to obtain wake flow field data.

[0049] Specifically, such as Figure 2 As shown, step S201 above includes:

[0050] Step S2011: Select simulation software based on computational fluid dynamics (CFD). Specifically, CFD mainly uses two theories for calculation: potential flow theory and viscosity theory. Potential flow theory focuses on quickly calculating the macroscopic characteristics of the flow field, such as velocity potential and pressure distribution. It is suitable for initially capturing the overall flow field trend of floating wind turbines under the action of wind and waves. It has high computational efficiency but limited accuracy. Viscous theory is based on the Navier-Stokes equations and can solve viscous flow. It can output flow field data (such as velocity, pressure, turbulence intensity, etc.) that includes details such as turbulence and wake distortion. It has high accuracy but high computational cost and time consumption.

[0051] When selecting simulation software, one can choose software based on potential flow theory, such as OpenFAST, or software based on viscosity theory, such as SOWFA and STAR-CCM+. Software based on potential flow theory can also be called potential flow software, and software based on viscosity theory can also be called CFD software.

[0052] Step S2012 involves using the aforementioned simulation software to model the floating wind turbines, wind farm layout, and parameters under different inflow conditions. The simulation software can simulate various environmental conditions; specifically, different inflow conditions such as wind speed, wind direction, turbulence intensity, wave period, and current velocity can be set in the software to cover typical operating conditions in complex marine environments. Furthermore, for wind farms formed by multiple floating wind turbines, parameters such as turbine spacing and arrangement also need to be set.

[0053] In addition, when performing modeling, it is also necessary to determine the wake spread rate. The formula for calculating the wake spread rate is:

[0054] d w =D+2k w

[0055] In the formula, D is the diameter of the wind turbine, and k w `dw` is the wake spread rate, used to describe the linear spread of the wake. `dw` is a linear coefficient describing the wake, specifically representing the rate of change of wake width with distance `x`. It is a specific numerical value used to describe the change in wake width at a specific location `x`. Using `dw`, the wake width at any location `x` can be calculated. This is crucial for understanding how the wake spreads downstream, especially in wind farm layout optimization, where the wake width at different locations needs to be known to assess its impact on downstream wind turbines. `kw` is a proportionality coefficient used to describe the linear relationship between wake width and distance `x`. It is a constant representing the change in wake width per unit distance. This constant is set to 0.075 for onshore wind farms and 0.050 for offshore wind farms.

[0056] Step S2013 involves determining the numerical solution method and boundary conditions, and refining the mesh in the wake region. Specifically, different numerical solution methods and boundary conditions can be used depending on the simulation software. For example, for potential flow software, potential flow theory can be used to simplify the flow equations, and the boundary conditions focus on the interaction between the fan and the fluid. For CFD software, the Navier-Stokes equations are solved, and a suitable turbulence model (such as the k-ε model) is selected. Boundary conditions include inlet velocity profiles, outlet free flow, and no-slip conditions on the walls to ensure the accuracy of the flow field's physical properties. In addition, the core wake region (such as the area 10-20 times the impeller diameter downstream of the fan) is locally refined to capture details such as wake velocity deficit and turbulent vortices. For non-wake regions, a coarser mesh can be used to balance computational efficiency and accuracy.

[0057] Step S2014 involves correcting the model parameters obtained from modeling and solving for the aerodynamic performance to obtain wake flow field data. Specifically, when simulating the model constructed using the above steps with numerical solution methods and corresponding boundary conditions, the model parameters can be adjusted based on the comparison between some measured data and simulation results to reflect the actual evolution process. After adjustment, the aerodynamic performance of the floating wind turbine and wind farm is solved using software to obtain high-fidelity, high-precision wake flow field data.

[0058] Step S202: Based on the wake flow field data, the indicator function representing the wake region is replaced with a smoothing step function. The wake region is represented by the interval [a, b). The indicator function indicates whether any point is within the wake region; if it is within the wake region, it returns 1; otherwise, it returns 0. The smoothing step function includes intervals [ad, a+d], [a+d, bd], [bd, b+d], and (-∞, ad) and (b+d, +∞). Within the interval [ad, a+d], the smoothing step function smoothly increases from 0 to 1; within the interval [a+d, bd], the smoothing step function returns 1; within the interval [bd, b+d], the smoothing step function smoothly decreases from 1 to 1; and within the intervals (-∞, ad) and (b+d, +∞), the smoothing step function returns 0. Here, d represents the standard deviation of the Gaussian distribution of the distance from any point in the wake region to the starting point of the wake region.

[0059] Specifically, the smoothing step function introduces a width parameter `d` to control the width of the transition region for the wake spread rate, based on the original interval. As `d` approaches 0, the approximation becomes more accurate. For the smoothing step function, a value of 1 indicates the area entirely within the wake region. A value of 0 indicates the area entirely outside the wake region. Areas with values ​​between 0 and 1 represent the transition region from outside to within the wake region. This transition region is introduced to avoid discontinuities, not to indicate that these points are within the wake region.

[0060] Step S203: Based on the adjusted wake flow field data in the wake region, determine the wake velocity deficit expression using a symbolic regression model based on particle swarm optimization algorithm.

[0061] Specifically, step S203 includes:

[0062] Step S2031: Obtain model training data based on the adjusted wake flow field data of the wake region. The model training data includes floating wind turbine arrangement parameters, amplitude, frequency, and normalized velocity deficit. Specifically, the arrangement parameters include parameters such as the spacing and arrangement of the wind turbines, and the amplitude and frequency include the difference between the maximum and minimum wind speeds and the frequency of wind turbine changes. The velocity deficit at a point in the wake follows a Gaussian formula, expressed as follows:

[0063]

[0064] In the formula, U represents the velocity deficit at a point in the wake, that is, the difference between the actual velocity at that point and the free-flow velocity (the velocity when there is no obstruction); U ∞ Let be the free-flow velocity; y and z be the vertical and horizontal distances from a point in the wake to the wake originating point, respectively; y0 and z0 are the locations where the velocity deficit reaches its maximum, typically directly below and in front of the wake originating point; σ is the standard deviation of the Gaussian distribution, controlling the width of the velocity deficit distribution. Based on this expression for the velocity deficit and combined with wake flow field data, the normalized velocity deficit can be determined. Furthermore, the width parameter d introduced in the smoothing step function above can be used to replace σ. y and σ x .

[0065] By using floating wind turbine layout parameters, amplitude, frequency, and normalized velocity loss as training data for the model, the model can learn the relationship between the input parameters and the normalized velocity loss, generating a mathematical expression to describe this relationship. This expression can be used to predict the normalized velocity loss under given wind turbine layout and environmental conditions.

[0066] Step S2032: Construct the search space of the symbolic regression model and generate an initial expression as particles for particle swarm optimization; specifically, based on the above expression, when constructing the search space participating in the symbolic regression model, the operator set is defined as (+, -, ×, ÷, exp, sin, ... Based on the aforementioned expression and the selection of basic parameters in the symbolic space, the function expression of the tail flow is obtained through training the model using symbolic regression. By constructing the hyperparameter space, the optimizer (Adam, SGD, RMSprop, etc.), the number of network layers, and the number of neurons are selected. Based on the training effect, the basic training framework is obtained.

[0067] After obtaining the basic training framework, error analysis is performed based on the numerical simulation results and model prediction results. The chosen metric, Mean Squared Error (MSE), is a commonly used indicator to evaluate the difference between model predictions and actual observations. It calculates the average of the squared differences between the predicted and actual values, and is used to measure the accuracy of the model's predictions. The smaller the MSE, the smaller the difference between the model's predictions and the actual values, and the better the model's performance. The formula is as follows:

[0068]

[0069] In the formula, N is the sample size; y i This is the i-th observation, i.e., the actual observed value; Let be the i-th observation, which is the value predicted by the symbolic regression model.

[0070] Step S2033 involves encoding, fitness evaluation, and optimization of the particles to determine the wake velocity deficit expression. Specifically, as follows: Figure 3 As shown, the particle swarm optimization algorithm's process includes initializing the particle swarm, then initializing individual and global maxima, and finally setting the number of iterations. Afterward, it enters the k+1 iteration stage, checking if the required number of iterations has been reached. If so, the optimal function expression is output and the process ends. If not, the particle velocity and position are updated, then the particle fitness is calculated. Next, it is checked whether the particle fitness is better than the individual maxima; if so, the individual maxima are updated. Then, it is checked whether the particle fitness is better than the global maxima; if so, the global maxima are updated. After completing these operations, the process returns to the k+1 iteration step, repeating this cycle until the required number of iterations is reached, finally outputting the optimal function expression and ending the process.

[0071] Specifically, in this embodiment, the flow field of the particle swarm optimization algorithm includes: treating the initial expression output by the symbolic regression model as a particle, encoding the particle using a gene expression, and encoding the syntax tree as a linear string to avoid code bloat, resulting in a simple and easy-to-implement structure. Then, a set of particles is randomly generated as initial candidate solutions.

[0072] Then, for each particle, its fitness function value is calculated. The fitness function can be the wake deficit rate η or other indicators that quantify wake loss. When the wake deficit rate is used as the fitness function, it is calculated using the following formula:

[0073]

[0074] In the formula, v wake The velocity v in the wake region represents the flow field velocity. streamThe free-flow velocity upstream is represented by the above formula. The wake loss rate of a single floating wind turbine can be calculated using this formula. Then, the wake loss rates η of all turbines are averaged or weighted summed to obtain the overall wake loss. Furthermore, when evaluating the fitness of particles, the minimum wake loss under the heave velocity ratio and yaw rate is used as a guide, specifically expressed by the following formula:

[0075]

[0076] In the formula, U max denoted as , where z is the maximum wind speed within the impeller sweep surface, and z is the heave displacement. For the yaw angle, U ∞ The upstream wind speed is the source of the interference from the primary wind turbine, and dz / dt is the heave velocity. Let be the yaw rate. The following function reflects the relationship between the wake width and the cumulative heave displacement, thus characterizing the physical interpretability of incorporating the particle swarm optimization algorithm:

[0077] σ y =0.5D + 0.1∫0 t |dz / dt|dt

[0078] In the formula, σ y The wake half-width determines the range of influence of the wake, and D is the diameter of the fan impeller.

[0079] After calculating the fitness value of each particle, the best-performing particles are selected. Then, based on the individual optimal solution (the position corresponding to the particle's own historical best fitness) and the global optimal solution (the position corresponding to the best fitness among all particles), the particle velocity is updated using the following formula:

[0080]

[0081] The particle structure (function expression) is adjusted according to the new velocity, and the position is updated using the following formula:

[0082]

[0083] Where ω is the inertia weight, balancing global and local search capabilities; c1 and c2 are acceleration constants; r1 and r2 are random numbers between [0,1]. and These are the current velocity and position of particle i. It is the individual extreme value of particle i. It is the global extremum of the group.

[0084] By continuously adjusting the position and velocity of particles, the particle optimization algorithm can effectively search for the optimal solution, fit the dataset, and find the best mathematical expression and parameter values. To avoid getting trapped in local solutions, a mutation mechanism similar to that of a genetic algorithm is introduced, and a mutation operator is designed. The mutation method is as follows: a random integer is generated as the number of mutation nodes, the mutation node position is randomly determined in the encoding region, and then the node is mutated. When the mutation position is in the first segment of the encoding region, it is randomly replaced with a function, operator, or terminator; when it is in the last segment of the encoding region, it is randomly replaced with a terminator.

[0085] Step S204: Evaluate the wake of the floating wind farm based on the wake velocity loss expression. During the evaluation, the wake loss under different blade deformation conditions can be assessed based on the wake velocity loss expression. Specifically, by comparing the time-averaged wake velocity loss distribution maps in the hub height plane under different blade deformation conditions, it is found that blade deformation has a relatively small impact on the wake velocity of the floating wind turbine. After the inflow wind passes through the rotor, a low-speed wake region is formed, and the wake velocity loss decreases slightly when the blades are deformed.

[0086] Step S205: Quantify wake loss based on the power output of the wind farm and the power output of a single floating wind turbine. Specifically, wake loss leads to energy loss, which is particularly significant below the rated wind speed of the turbine. Therefore, wake loss can be modeled using the wind farm power model and the equivalent model of a single wind turbine, and expressed by the following formula:

[0087] P max (ws,wd,ti,wd var ,ρ)=J×P WT (ws,ti,wd var ,ρ)-P(s,wd,ti,wd var ,ρ)

[0088] In the formula, ws, wd, ti, wd var ρ and ρ represent the inflow wind speed, wind direction, flow intensity, wind direction variance, and air density of the wind farm, respectively. J represents the scaling factor, used to describe the proportion of power loss due to the wake effect. Pmax represents the maximum possible power output of the wind farm under given conditions. This is the power output under ideal conditions, assuming no wake loss. WT This represents the power output of a single wind turbine. This is the power output of a single turbine under given wind speed, turbulence intensity, and other conditions. P represents the actual power output of the wind farm. This is the actual power output after considering wake losses. This formula allows comparison of the actual power output of a wind farm with the power output under ideal conditions, thereby quantifying wake losses.

[0089] The wake assessment method for floating wind farms provided by this invention has the following advantages:

[0090] (1) Establish a multi-objective optimization integration framework to achieve high computational efficiency.

[0091] Combining the high precision of numerical simulation with the high efficiency of data-driven approaches, this invention constructs a wake model for multiple environmental factors at sea, based on particle swarm optimization and symbolic regression models, and constrained by physical equations, to simulate the impact of wake disturbances in floating wind farms.

[0092] (2) The ability to apply wind farms in complex environments and to achieve model optimization.

[0093] Offshore measurement data is difficult to obtain and costly. This invention combines numerical simulation data of the aerodynamic performance of wind farms under different environmental conditions obtained by CFD methods for model training and optimization.

[0094] (3) Dynamically adjust the model complexity to balance the fit and generalization ability.

[0095] This invention enables rapid simulation and prediction of the wake of floating wind turbines. Traditionally, CFD is used for integrated numerical calculations of wind turbines, which is time-consuming and costly. This invention uses a particle swarm optimization algorithm to compare multiple equations and select the optimal symbolic regression equation, enabling rapid prediction of key areas of concern in engineering projects.

[0096] As a specific application embodiment of the present invention, such as Figure 4 As shown, the wake assessment method for this floating wind farm can be implemented through the following process: First, acquire data such as the location information of each floating wind turbine in the floating wind farm. Then, perform numerical modeling and simulation calculations based on potential flow and CFD software to obtain high-precision wake field data. Next, construct a symbolic regression expression for the turbine environment and design parameters based on a symbolic regression model, serving as the floating wind turbine wake model. Finally, use a particle swarm optimization algorithm to find the optimal wake velocity loss expression. Based on this expression, predict and assess the overall wind farm wake loss.

[0097] This embodiment also provides a wake assessment device for a floating wind farm, which is used to implement the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0098] This embodiment provides a wake assessment device for floating wind farms, such as... Figure 5 As shown, it includes:

[0099] Simulation module 51 is used to simulate the operation of the floating fan using computational fluid dynamics methods to obtain wake flow field data;

[0100] The function replacement module 52 is used to replace the indicator function characterizing the wake region with a smoothing step function based on the wake flow field data.

[0101] The expression determination module 53 is used to determine the wake velocity deficit expression based on the wake flow field data after the wake region is adjusted, using a symbolic regression model based on particle swarm optimization algorithm.

[0102] Evaluation module 54 is used to evaluate the wake of a floating wind farm based on the wake velocity deficit expression.

[0103] Further functional descriptions of the above modules are the same as those in the corresponding embodiments described above, and will not be repeated here.

[0104] This invention also provides a computer device having the above-described features. Figure 5 The wake assessment device for the floating wind farm shown.

[0105] Please see Figure 6 , Figure 6 This is a schematic diagram of the structure of a computer device provided in an optional embodiment of the present invention, such as... Figure 6 As shown, the computer device includes one or more processors 10, memory 20, and interfaces for connecting the components, including high-speed interfaces and low-speed interfaces. The components communicate with each other via different buses and can be mounted on a common motherboard or otherwise installed as needed. The processors can process instructions executed within the computer device, including instructions stored in or on memory to display graphical information of a GUI on external input / output devices (such as display devices coupled to the interfaces). In some alternative implementations, multiple processors and / or multiple buses can be used with multiple memories and multiple memory modules, if desired. Similarly, multiple computer devices can be connected, each providing some of the necessary operations (e.g., as a server array, a group of blade servers, or a multiprocessor system). Figure 6 Take a processor 10 as an example.

[0106] Processor 10 may be a central processing unit, a network processor, or a combination thereof. Processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The programmable logic device may be a complex programmable logic device (CAMP), a field-programmable gate array (FPGA), a general-purpose array logic (GDA), or any combination thereof.

[0107] The memory 20 stores instructions executable by at least one processor 10 to cause at least one processor 10 to perform the method shown in the above embodiments.

[0108] The memory 20 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the computer device as shown by a landing page for an app. Furthermore, the memory 20 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some alternative embodiments, the memory 20 may optionally include memory remotely located relative to the processor 10, which can be connected to the computer device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0109] The memory 20 may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as flash memory, hard disk or solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0110] The computer device also includes a communication interface 30 for communicating with other devices or communication networks.

[0111] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code, which, when accessed and executed by the computer, processor, or hardware, implements the methods shown in the above embodiments.

[0112] A portion of this invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to the invention through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0113] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for assessing the wake of a floating wind farm, characterized in that, The method includes: The operation of the floating fan was simulated using computational fluid dynamics to obtain wake flow field data; Based on the wake flow field data, the indicator function characterizing the wake region is replaced with a smoothing step function; Based on the adjusted wake flow field data in the wake region, a symbolic regression model based on particle swarm optimization algorithm is used to determine the wake velocity deficit expression. The wake of a floating wind farm is evaluated based on the aforementioned wake velocity deficit expression.

2. The method according to claim 1, characterized in that, The operation of the floating wind turbine was simulated using computational fluid dynamics methods to obtain wake flow field data, including: Screening simulation software based on computational fluid dynamics methods; The simulation software was used to model floating wind turbines, wind farm layouts, and parameters under different inflow conditions; Determine the numerical solution method and boundary conditions, and refine the mesh in the wake region; The model parameters obtained from modeling are corrected, and the aerodynamic performance is solved to obtain the wake flow field data.

3. The method according to claim 1, characterized in that, The wake region is represented by the interval [a, b). The indicator function indicates whether any point is within the wake region. If it is within the wake region, it returns 1; otherwise, it returns 0. The smoothing step function includes intervals [ad, a+d], [a+d, bd], [bd, b+d], and (-∞, ad) and (b+d, +∞). Within the interval [ad, a+d], the smoothing step function smoothly increases from 0 to 1. Within the interval [a+d, bd], the smoothing step function returns to 1. Within the interval [bd, b+d], the smoothing step function smoothly decreases from 1 to 1. Within the intervals (-∞, ad) and (b+d, +∞), the smoothing step function returns to 0. Here, d represents the standard deviation of the Gaussian distribution of the distance from any point in the wake region to the starting point of the wake region.

4. The method according to claim 1, characterized in that, Based on the adjusted wake flow field data, a symbolic regression model based on particle swarm optimization is used to determine the wake velocity deficit expression, including: Model training data is obtained based on the wake flow field data after adjustment of the wake region. The model training data includes floating wind turbine layout parameters, amplitude, frequency, and normalized velocity loss. Construct the search space for the symbolic regression model and generate initial expressions as particles for particle swarm optimization; The particles are encoded, their fitness is evaluated, and optimized to determine the wake velocity deficit expression.

5. The method according to claim 4, characterized in that, When evaluating the fitness of particles, the minimum wake loss under the heave velocity ratio and yaw angular velocity is used as the guide, and the minimum wake loss rate is used as the fitness value.

6. The method according to claim 5, characterized in that, The relationship between the heave velocity ratio, yaw rate, and wake loss is expressed by the following formula: In the formula, U max denoted as , where z is the maximum wind speed within the impeller sweep surface, and z is the heave displacement. For the yaw angle, U ∞ The upstream wind speed is the source of the interference from the primary wind turbine, and dz / dt is the heave velocity. This is the yaw rate.

7. The method according to claim 1, characterized in that, The method for evaluating the wake of a floating wind farm based on the wake velocity deficit expression includes: evaluating the wake loss under different blade deformation conditions based on the wake velocity deficit expression; the method further includes: Quantify wake loss based on the power output of the wind farm and the power output of a single floating wind turbine.

8. A wake assessment device for a floating wind farm, characterized in that, The device includes: The simulation module is used to simulate the operation of the floating wind turbine using computational fluid dynamics methods to obtain wake flow field data; The function replacement module is used to replace the indicator function characterizing the wake region with a smoothing step function based on the wake flow field data; The expression determination module is used to determine the wake velocity deficit expression based on the wake flow field data after adjustment of the wake region using a symbolic regression model based on particle swarm optimization algorithm. An evaluation module is used to evaluate the wake of a floating wind farm based on the wake velocity deficit expression.

9. A computer device, characterized in that, include: A memory and a processor are communicatively connected, the memory stores computer instructions, and the processor executes the computer instructions to perform the wake assessment method for a floating wind farm as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to execute the wake assessment method for the floating wind farm as described in any one of claims 1 to 7.

Images