A method for parameter tuning of a tuned mass damper for a floating wind turbine

Abstract

The present application relates to offshore floating wind turbine technical field, disclose a kind of for the parameter tuning method of tuned mass damper of floating wind turbine, comprising: the rigid-flexible coupling multibody dynamics model of whole coupling numerical model and drive train is constructed;Parameter identification is carried out;The frequency domain dynamics model of first tuned mass damper arranged in cabin and second tuned mass damper arranged in floating platform is constructed;Establish parameter optimization problem;Global optimization is carried out using pattern search algorithm, then local fine optimization is carried out using simplex search algorithm, and optimal parameter combination is obtained.The load characteristics in the drive train and axial displacement response are included in the optimization index, and the multi-degree-of-freedom coupled dynamics modeling of double tuned mass damper and hybrid algorithm precise optimization are used to solve the problem that the existing technology cannot effectively alleviate the complexity and extreme load of the drive train due to the arrangement limitation of TMD and the deviation of parameter tuning target, which easily causes early damage of component fatigue.

Description

Technical Field

[0001] This invention relates to the field of offshore floating wind turbine technology, and specifically to a method for tuning the parameters of a tuned mass damper for a floating wind turbine. Background Technology

[0002] Floating wind turbines are core equipment for deep-sea wind power development. Their drive system, as a crucial link in energy transfer, directly affects the unit's power generation efficiency and service life. Unlike stationary wind turbines, floating turbines are subjected to the coupled effects of multiple random loads such as wind, waves, and currents during operation. The multi-degree-of-freedom motion of the floating platform alters the input characteristics of aerodynamic loads, transmitting significant axial dynamic loads to the drive system through the tower and rotor. This causes critical components such as gears and bearings within the drive system to experience severe alternating stresses, making them highly susceptible to fatigue cracks and seriously threatening the unit's operational safety. Due to the harsh environment of deep-sea areas, once the drive system is damaged, the repair window is short, operating costs are extremely high, and prolonged power generation shutdowns result in substantial economic losses.

[0003] Tuned mass dampers (TMDs), as a classic passive vibration reduction device, have been widely used in vibration control of various engineering structures due to their simple structure, lack of external power requirements, and high reliability. In the field of floating wind turbines, existing technologies have also proposed placing TMDs at the nacelle or platform to suppress the macroscopic vibration of the entire turbine. For example, existing research has established simplified dynamic models of floating wind turbines that include nacelle TMDs or platform TMDs, and uses intelligent optimization algorithms such as genetic algorithms and artificial fish swarm algorithms to optimize the stiffness, damping, and mass parameters of the TMDs, with the optimization objective of minimizing the standard deviation of tower top displacement or platform roll angle, in order to reduce the dynamic response of the entire turbine.

[0004] However, the aforementioned existing technologies all have significant limitations. On the one hand, existing TMD parameter tuning methods all aim to suppress the overall macroscopic vibration of the wind turbine, such as suppressing the deformation of the tower and the pitching of the floating platform. Their optimization indices are all structural responses at the overall turbine level, without treating the transmission system, a core and vulnerable internal component, as an independent optimization object. Therefore, they cannot specifically alleviate the dynamic loads borne by the transmission system. On the other hand, existing methods typically simplify the wind turbine into a lumped mass model with a few degrees of freedom at the modeling level, failing to establish a high-fidelity dynamic model that can simultaneously characterize the global motion of the entire turbine and the load transmission within the transmission system. As a result, although the optimized TMD parameters can reduce the overall turbine vibration amplitude, they cannot guarantee the effective reduction of the load within the transmission system, and may even exacerbate fatigue damage to specific components of the transmission system due to the redistribution of vibration energy.

[0005] In summary, how to establish a TMD parameter tuning method with the goal of directional load reduction of the transmission system, while fully considering the coupled dynamic characteristics of the floating wind turbine and its transmission system, has become an urgent technical problem to be solved in this field. Summary of the Invention

[0006] To address the technical problems in existing technologies where the tuning target of TMD parameters is limited to macroscopic vibration suppression of the entire machine and fails to target load reduction based on the internal load characteristics of the transmission system, this invention proposes a tuning method for the parameters of a tuned mass damper for load reduction in the transmission system of a floating wind turbine. The technical solution is as follows: A method for tuning parameters of a tuned mass damper for a floating wind turbine, characterized by comprising the following steps: S1. Construct a dynamic model of the entire floating wind turbine and its transmission system: Establish a fully coupled numerical model of the entire turbine by equating the transmission system with a single-degree-of-freedom spring-damped system, and establish a rigid-flexible coupled multibody dynamic model of the transmission system separately. S2. Based on the fully coupled numerical model of the whole machine, construct a linear time-domain dynamic model of the floating wind turbine; S3. Using the dynamic response data output by the fully coupled numerical model of the whole machine, the unknown parameters in the linear time-domain dynamic model are identified to obtain the parameter-calibrated linear time-domain dynamic model. S4. Based on the linear time-domain dynamic model calibrated by parameters, a frequency-domain dynamic model including a tuned mass damper is constructed. The tuned mass damper includes a first tuned mass damper arranged in the cabin and a second tuned mass damper arranged on the floating platform. S5. Using the stiffness and damping parameters of the first and second tuned mass dampers as optimization variables, and minimizing the H2 norm of the axial displacement of the wind turbine and transmission system under wind and wave load input as the objective function, a parameter optimization problem is established. S6. Use a pattern search algorithm to perform global optimization on the parameter optimization problem and obtain the region where the global optimal solution is located; S7. The simplex search algorithm is used to perform local fine optimization in the region where the global optimal solution is located, so as to obtain the optimal parameter combination of the first tuned mass damper and the second tuned mass damper.

[0007] Furthermore, it also includes: S8, backfilling the optimal parameter combination into the fully coupled numerical model of the whole machine for fully coupled time-domain simulation, extracting the input load of the transmission system and the motion data of the platform as boundary conditions, inputting them into the rigid-flexible coupled multibody dynamics model of the transmission system to solve the load response of the internal key components, and using the fatigue damage degree of the key components of the transmission system as the verification index of the optimal parameter combination.

[0008] Furthermore, the parameter identification in step S3 includes: identifying unknown parameters in the linear time-domain dynamic model with the goal of minimizing the root mean square error between the tower deformation response output by the fully coupled numerical model of the whole machine and the tower deformation response output by the linear time-domain dynamic model.

[0009] Furthermore, the frequency domain dynamic model in step S4 is a five-degree-of-freedom frequency domain dynamic model, which is composed of the three-degree-of-freedom dynamic model of the entire floating wind turbine superimposed with the single-degree-of-freedom motion equations of the first tuned mass damper and the second tuned mass damper.

[0010] Further, in step S5, the objective function is the sum of the H2 norm of the wind load transfer function and the H2 norm of the wave load transfer function, expressed as:

[0011] in, The transfer function for the axial displacement of the wind turbine and transmission system caused by wind load; The transfer function for the axial displacement of the wind turbine and transmission system caused by wave load.

[0012] Furthermore, the global optimization of the pattern search algorithm in step S6 includes: generating an initial search starting point in the constraint space, setting an initial grid size and step size, and calculating the objective function value of the trial point in each exploration direction; if the objective function value decreases, the grid step size is increased and the trial point is used as the new iteration center point; if the objective function value does not decrease, the grid step size is reduced; when the grid step size is lower than the preset switching threshold, the pattern search ends and the current optimal point is output.

[0013] Furthermore, the local fine optimization of the simplex search algorithm in step S7 includes: taking the current optimal point output in the pattern search stage as the first vertex of the initial simplex, constructing the initial simplex in the neighborhood of the current optimal point; calculating the objective function values ​​of all vertices and sorting them according to their quality; performing reflection, expansion, contraction, or edge shrinking operations on the worst point to generate candidate vertices; replacing the candidate vertex if its objective function value is better than the current worst point; iterating in a loop until the convergence criterion is met, and outputting the final global optimal point.

[0014] Furthermore, the evaluation of the fatigue damage degree of key components of the transmission system in step S8 includes: using dynamic equivalent load conversion to obtain the number of cycles of key bearings in each load range, and combining linear cumulative damage theory and wind field operating probability to calculate the long-term cumulative fatigue damage degree of key components of the transmission system; if the fatigue damage degree meets the preset engineering threshold, the optimal parameter combination is output; otherwise, the optimal parameter combination is used as a new initial point to return to step S6 for re-optimization.

[0015] Furthermore, step S5 also includes: setting upper and lower bound constraints for the optimization variables based on the overall load-bearing capacity of the floating wind turbine and the actual layout space inside the nacelle and platform.

[0016] Compared with the prior art, the advantages and positive effects of the present invention are as follows: This invention innovatively incorporates the internal load characteristics and axial displacement response of the transmission system into the optimization index system. It constructs a multi-degree-of-freedom coupled dynamic model including dual-tuned mass dampers for the nacelle and floating platform, and combines pattern search and simplex search algorithms to achieve precise parameter optimization. This method is particularly suitable for the design of passive vibration reduction systems for floating wind turbines under complex multi-coupled excitation conditions in deep-sea environments, where the system inertia is large, the degree-of-freedom coupling is strong, and the fatigue life requirements for the transmission system are extremely stringent. It effectively solves the technical problem of existing technologies, which, due to limitations in the arrangement of tuned mass dampers and deviations in parameter tuning targets, are unable to directionally alleviate complex extreme loads on the transmission system, thus leading to early fatigue damage of components.

[0017] This invention employs a hybrid optimization strategy combining pattern search and simplex search algorithms. The pattern search algorithm globally explores the objective function within the constraint space using a grid search mechanism, quickly identifying the region containing the optimal parameter combination, effectively overcoming the tendency of single intelligent optimization algorithms to get trapped in local optima. Subsequently, the simplex search algorithm performs a fine approximation within this region using derivative-free geometric transformations, rapidly converging to the minimum point of the objective function. The two algorithms sequentially undertake the different tasks of global coarse exploration and local fine search, significantly improving the computational efficiency of parameter optimization while ensuring solution accuracy. This results in the obtained stiffness and damping parameter combination of the tuned mass damper being closer to the theoretical optimal solution, thereby maximizing the load reduction effect of the transmission system. Attached Figure Description

[0018] The invention will now be further described with reference to the accompanying drawings.

[0019] Figure 1 This is a circuit diagram of the method provided in Embodiment 1 of the present invention; Figure 2 This is a schematic diagram of the model established in the embodiment of the present invention, wherein (a) is a schematic diagram of the fully coupled numerical model of the floating wind turbine, and (b) is a schematic diagram of the rigid-flexible coupled multibody dynamics model of the transmission system; Figure 3 This is a comparison diagram of the amplitude-frequency characteristics of the transfer function of the wind excitation and wave excitation to the axial displacement of the wind turbine and the transmission system before and after parameter tuning in the embodiment of the present invention. Among them, (a) is the wind excitation channel and (b) is the wave excitation channel. Figure 4This is a comparison diagram of the dynamic equivalent load distribution of the upwind and downwind main bearings before and after parameter tuning in an embodiment of the present invention, wherein (a) is the upwind main bearing and (b) is the downwind main bearing. Detailed Implementation

[0020] This invention will be further described in detail with reference to the accompanying drawings and the following specific embodiments. It should be understood that the specific embodiments described herein are only for explaining the invention and are not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of the invention described below can be combined with each other as long as they do not conflict with each other. This embodiment provides a method for tuning the parameters of a tuned mass damper for a floating wind turbine. The specific embodiments of the invention will be described in detail below according to the method steps.

[0021] Example, reference Figure 1 .

[0022] I. Constructing a multi-level dynamic model of the floating wind turbine and its transmission system In this embodiment, step S1 constructs a multi-level dynamic model of the floating wind turbine and its transmission system, which includes two parts: constructing a fully coupled numerical model of the whole machine and constructing a rigid-flexible coupled multibody dynamic model of the transmission system.

[0023] (I) Constructing a fully coupled numerical model of the whole machine This embodiment establishes a fully coupled numerical model of the floating wind turbine based on a numerical simulation platform. Specifically, based on a joint simulation architecture combining MATLAB / Simulink with Simpack and OpenFAST, the model defines in detail the mass, stiffness, damping, and moment of inertia distribution parameters of each key structure of the floating wind turbine. In the model, the foundation floating platform is defined as a six-degree-of-freedom rigid body, capable of characterizing its sway, heave, roll, pitch, and yaw motions in space; the tower and blades are modeled as multimodal elastic flexible bodies, and their structural elastic deformation is accounted for through modal superposition. To ensure the efficiency of solving the entire machine, the transmission system is simplified to a single-degree-of-freedom spring-damped system in this model.

[0024] Regarding environmental loads, based on actual deep-sea operating conditions, a three-dimensional random turbulent wind field time history is generated using the Kaimal wind speed spectrum or IEC standard specifications. The nonlinear aerodynamic loads on the wind turbine and blades during rotation are calculated using blade element momentum theory. Irregular wave excitation time histories are generated using the JONSWAP or Pierson-Moskowitz irregular wave spectrum, and potential flow theory is used to calculate the hydrostatic restoring force, added mass, radiation damping, and diffraction loads experienced by the foundation floating platform under wave action. The lumped mass method, combined with Morrison's equations, is used to simulate the dynamic tension, gravity, buoyancy, and seabed contact characteristics of the subsea mooring cables underwater, thereby setting the subsea mooring constraints for the floating platform system. Furthermore, an integrated closed-loop controller is implemented to execute variable-speed generator torque control and collective pitch adjustment control strategies to stabilize generator power and regulate wind turbine speed.

[0025] (II) Constructing a rigid-flexible coupled multibody dynamics model of the transmission system This embodiment establishes a high-fidelity rigid-flexible coupled multibody dynamics model of the transmission system based on the Simpack multibody dynamics simulation platform. Specifically, the main shaft, planetary carrier, and gearbox shaft are treated as flexible bodies, and the Craig-Bampton method is used for modal order reduction to account for the impact of large structural deformations on the dynamic characteristics of the transmission system. The hub and base plate are treated as concentrated mass rigid bodies, and the degrees of freedom between components are constrained by kinematic pairs. The gear pairs adopt a multi-tooth meshing model, and the comprehensive gear force of each contact tooth is calculated and superimposed based on the tooth profile penetration. The main bearing, gearbox bearing, and generator coupling are all represented as spring-damped components using linearized stiffness and damping matrices that do not consider cross-coupling. The gearbox torque arm is equivalent to a linear spring-damped model to provide boundary constraints. In addition, a proportional-integral controller is integrated to adjust the generator torque in a closed loop based on the generator speed.

[0026] The two models described above have different functions in subsequent steps: the fully coupled whole machine model is used to generate the dynamic response of the whole machine under environmental excitation, while the multibody transmission model further calculates the fine loads of key internal components based on this response, providing a quantitative evaluation method for parameter tuning with the goal of reducing the load on the transmission system.

[0027] Figure 2 (a) in the figure is the fully coupled numerical model of the whole machine established on the whole machine simulation platform according to (a) above. Figure 2(b) in the figure is the rigid-flexible coupled multibody dynamics model of the transmission system established on the multibody dynamics simulation platform according to (ii) above. As shown in the figure, the former uses blades, hubs, nacelles, towers, floating foundation platforms, and mooring systems as modeling objects to solve the macroscopic dynamic response of the whole machine under the combined excitation of wind and waves; the latter uses components such as hubs, main shafts, main bearings, gearboxes, and base plates as modeling objects to solve the fine load response of key components inside the transmission system under the load and motion boundary conditions output by the former. Together, they constitute the hierarchical load transfer analysis architecture of this invention from the whole machine scale to the component scale, so that the load inside the transmission system is transformed from an unobservable quantity at the whole machine level into a computable and quantifiable optimization and verification index.

[0028] II. Constructing a simplified linear time-domain dynamic model for floating wind turbines Based on the above fully coupled numerical model of the whole machine, a simplified linear time-domain dynamic model of the floating wind turbine is constructed.

[0029] Since the fully coupled numerical model of the whole machine established in the above steps has a large degree of freedom and high computational cost, it is not suitable for direct iterative optimization of TMD parameters. Therefore, this invention constructs a simplified linear time-domain dynamic model that can significantly reduce computational cost while retaining the main characteristics of the axial displacement response of the transmission system.

[0030] This embodiment considers three degrees of freedom: the wind turbine tower angle, the pitch of the floating platform, and the axial translation of the wind turbine and the transmission system. A simplified linear time-domain dynamic model of the floating wind turbine is established, and its dynamic equations are as follows:

[0031] in, , , These represent the masses of the floating platform, tower, and wind turbine drive system unit, respectively. and These represent the moments of inertia of the floating platform and the tower, respectively. , , These represent the center of mass of the floating platform, the center of mass of the tower, and the distances from the engine room to the connection points of the tower and the floating platform, respectively. and These represent the stiffness and damping coefficients of the platform's pitching motion, respectively. and These represent the stiffness and damping coefficients of the axial displacement between the wind turbine drive unit and the nacelle, respectively. and These represent the stiffness and damping coefficients of the tower under front and rear deformations, respectively. , , These represent the tower angle, the platform pitch, and the axial translational displacement of the wind turbine and transmission system, respectively. This indicates the offset of the center of mass of the wind turbine drive unit from the center line of the tower. and These represent wave excitation and wind excitation, respectively.

[0032] This simplified model retains the dynamic characteristics directly related to the axial displacement of the transmission system in the form of three-degree-of-freedom lumped mass parameters, providing an analytically expressible mathematical model for subsequent parameter identification and frequency domain modeling.

[0033] This invention constructs a fully coupled numerical model of the entire machine while simultaneously establishing a separate rigid-flexible coupled multibody dynamics model of the transmission system. By using the transmission system input loads and nacelle motion data obtained from the fully coupled time-domain simulation of the entire machine as boundary conditions input to the transmission system multibody model, it achieves hierarchical load transfer analysis decoupled from the entire machine and its components. This architecture transforms internal transmission system loads, such as gear meshing forces and bearing dynamic contact forces, which are currently unobservable at the whole-machine level, into calculable and quantifiable engineering indicators. This provides precise physical input for subsequent tuning of the parameters of the tuned mass damper, aimed at reducing transmission system fatigue.

[0034] III. Parameter Identification of Linear Dynamic Systems In this embodiment, the dynamic response data output by the fully coupled numerical model of the whole machine is used to identify the unknown parameters in the simplified linear time-domain dynamic model.

[0035] Since the simplified linear time-domain dynamic model established in the above steps contains several unknown parameters, including the platform's pitch stiffness... Floating platform pitch damping Moment of inertia of the floating platform Tower front and rear deformation stiffness Deformation damping of the tower front and rear These parameters are difficult to obtain directly through physical measurements and need to be extracted from high-fidelity models using parameter identification methods. Specifically, the pitch freedom degree of the floating platform in the fully coupled numerical model of the whole machine is set to 5°, and its free decay response curve is obtained. With the goal of minimizing the root mean square error between the tower's front-to-back deformation response and pitch response output by the fully coupled numerical model of the whole machine and the tower's front-to-back deformation response and pitch response output by the simplified linear time-domain dynamic model, a particle swarm optimization algorithm is used to identify the aforementioned unknown parameters. The root mean square error is expressed as:

[0036] in, Total data volume; The first output of the high-fidelity model One result; To simplify the output of the linear dynamics model One result.

[0037] Through iterative optimization, the dynamic response of the simplified linear time-domain dynamic model under the free attenuation condition of the floating platform's pitch is made consistent with that of the high-fidelity model, thus obtaining a parameter-calibrated simplified linear time-domain dynamic model, providing an accurate parameter basis for the subsequent establishment of the frequency domain model.

[0038] IV. Constructing a frequency domain dynamic model of a floating wind turbine and its vibration reduction system Based on a simplified linear time-domain dynamic model calibrated with parameters, a frequency-domain dynamic model incorporating a tuned mass damper is constructed. This tuned mass damper includes a first tuned mass damper located in the engine room and a second tuned mass damper located on the floating platform.

[0039] This step constructs a five-degree-of-freedom frequency domain dynamic model, which is composed of a three-degree-of-freedom dynamic model of the entire floating wind turbine superimposed with the single-degree-of-freedom motion equations of the first tuned mass damper and the second tuned mass damper.

[0040]

[0041] in, and These are the masses of the first tuned mass damper and the second tuned mass damper, respectively. and The displacements of the first tuned mass damper and the second tuned mass damper; and These are the transfer functions of the first tuned mass damper and the second tuned mass damper, respectively. This indicates the distance from the center of mass of the second tuned mass damper to the connection point between the tower and the floating platform.

[0042] This step transforms the time-domain dynamic model to the frequency domain, enabling the transmission relationship between wind and wave loads and the axial displacement of the transmission system to be explicitly expressed using a transfer function. This provides a mathematical foundation for the analytical calculation of the objective function and the rapid iteration of subsequent optimization algorithms.

[0043] V. Define optimization variables, objective function, and constraints. This embodiment defines the optimization variables, objective function, and constraints for TMD parameter optimization. The stiffness and damping parameters of the first and second tuned mass dampers, respectively located in the nacelle and on the floating platform, are used as optimization variables. The objective function is to minimize the H² norm of the axial displacement of the wind turbine and transmission system under wind and wave load input. The objective function is defined as follows:

[0044] in, The axial displacement of the wind turbine drive unit is input to the wind and wave loads. Sum of norms; The transfer function for the axial displacement of the wind turbine and transmission system caused by wind load; The transfer function for the axial displacement of the wind turbine and transmission system caused by wave loads.

[0045] This invention optimizes the axial displacement of the wind turbine and transmission system under wind and wave load input by minimizing the H2 norm. This norm characterizes the cumulative energy effect of the system's response across the entire frequency range and can serve as a quantitative indicator of the severity of transmission system load fluctuations. By setting this norm as the optimization objective, this invention precisely guides the tuning direction of the tuned mass damper from traditional macroscopic structural response indicators of the entire machine, namely tower top displacement and platform roll angle, to minimizing the total vibration energy input by wind and waves to the transmission system. This focuses the parameter configuration of the tuned mass damper on the core objective of the internal load response of the transmission system, fundamentally solving the technical problem of insufficient transmission system load relief due to target deviation in existing tuning strategies, and achieving directional reduction of the internal load of the transmission system.

[0046] VI. Global TMD Parameter Optimization Based on Pattern Search Algorithm In this embodiment, a pattern search algorithm is used to perform global optimization on the above parameter optimization problem in order to quickly identify the region where the optimal solution is located. The specific steps are as follows: The first step is to generate an initial search starting point X0 within the constrained space based on engineering experience, set an initial grid size Δ0, and generate a set of positive generation directions D={+e1, -e1, …, +e_n, -e_n} covering all dimensions, where e_i is the unit orthogonal vector in the i-th coordinate direction, n is the dimension of the parameter to be optimized, and the total number of directions M=2n.

[0047] The second step is to generate trial points in the exploration directions around the current center point X_k, call the objective function J defined in step five, and calculate the objective function value of each trial point; for trial points that exceed the constraint space, set their objective function value to infinity.

[0048] The third step is to determine if there is a trial point that causes the objective function J to decrease significantly from the current center point. If so, the current polling is considered successful. The trial point is then used as the new iteration center point X_{k+1}, and the grid step size Δ_{k+1} = 2·Δ_k is increased to accelerate the exploration.

[0049] Fourth step: If no point is found in all exploration directions that reduces the objective function J, the current polling is considered a failure. Keep the current center point position unchanged and reduce the grid step size Δ_{k+1} = 0.5·Δ_k.

[0050] Fifth, repeat steps two through four until the grid step size Δ_k is lower than the preset threshold Δ_th. At this point, it is considered that the neighborhood with the smaller objective function value has been located, the current optimal point X_best is output, and the global search ends.

[0051] This step utilizes the grid search mechanism of the pattern search algorithm to perform a coarse-to-fine search within the constrained space without requiring gradient information of the objective function. The initially larger grid step size allows the search to cross local extrema smaller than that step size, thereby reducing the sensitivity of the optimization results to the selection of the initial point and providing a more reliable initial point for subsequent local searches.

[0052] VII. Local TMD Parameter Optimization Based on Simplex Search Algorithm The simplex search algorithm is a direct search algorithm based on geometric transformations. It guides the search point sequence towards the minimum value of the objective function by constructing and iteratively updating the simplex geometry in the variable space. This invention passes the optimal point X_best output from the pattern search stage to the simplex search algorithm as the first vertex of its initial simplex, performing local fine-tuning in its neighborhood. The specific steps are as follows: The first step is to construct the initial simplex: Within the neighborhood of X_best, offset along each coordinate axis by a preset step size to generate an initial simplex geometry containing N+1 vertices, where N is the dimension of the optimization variable vector X. The remaining vertices of the initial simplex are generated by offsetting X_best along each coordinate axis by the preset step size.

[0053] The second step is vertex evaluation and sorting: calculate the objective function J value of all vertices of the simplex, sort them in ascending order of objective function value, and identify the current best point X_L, second worst point X_S, and worst point X_H.

[0054] The third step is geometric transformation iteration: Calculate the geometric center X_C = (1 / n)·Σ_{i≠H}X_i of all vertices except the worst point X_H. Based on the comparison between the reflection point and the objective function value of each vertex, perform a reflection operation X_R = X_C + α·(X_C-X_H) on the worst point. If the reflection point is worse than the second worst point X_S, perform a contraction operation X_K = X_C + β·(X_H - X_C) to generate new candidate vertices. Here, α, γ, and β are the reflection coefficient, expansion coefficient, and contraction coefficient, respectively. In this embodiment, α=1, γ=2, and β=0.5, respectively. If a vertex generated by the above geometric transformation exceeds the constraint space, its objective function value is set to infinity, making it unacceptable.

[0055] The fourth step is simplex update: If the objective function value of the candidate vertex is better than the current worst point X_H, then the candidate vertex replaces the worst point, completing the simplex update and transformation; otherwise, the edge shrinking operation is performed to move all vertices closer to the current best point X_L, that is, update according to X_i=X_L+δ·(X_i-X_L) (i≠L), where δ is the edge shrinking coefficient, and in this embodiment, δ=0.5 is taken.

[0056] The fifth step involves iterating through the above geometric transformation steps until the standard deviation of the objective function value at each vertex of the simplex is lower than the preset tolerance or the volume of the simplex is less than the threshold. At this point, the final optimal point X_best is output, which is the optimal combination of stiffness parameters and optimal damping parameters of the first and second tuned mass dampers that achieve the best load reduction effect of the floating wind turbine drive system.

[0057] This step leverages the advantage of the simplex search algorithm in rapidly approximating the minimum point through geometric transformation within a local region. It performs fine optimization within the neighborhood of the objective function value defined by the pattern search, enabling the final TMD parameter combination to minimize the H2 norm of the axial displacement of the transmission system.

[0058] Figure 3 To compare the frequency domain verification results of the parameter tuning effect of this invention, the method of obtaining the results is as follows: The optimal parameter combination obtained in this step is substituted into the five-degree-of-freedom frequency domain dynamic model established in step four to calculate the transfer function from wind load and wave load to the axial displacement of the wind turbine and transmission system. This transfer function is then compared with the transfer function obtained from the model calibrated in step three under the original configuration without a tuned mass damper. Figure 3 (a) in the diagram represents the wind excitation channel. Figure 3 (b) in the figure represents the wave excitation channel; the dashed line in the figure represents the original configuration, and the solid line represents the configuration after arranging the dual-tuned mass dampers with the optimal parameter combination. (Reference) Figure 3 It is known that the original configuration of the transfer function has sharp resonance peaks at the frequencies corresponding to the platform's pitch mode and the tower's first-order forward and backward bending modes. These two frequencies are the main channels for the axial transmission of wind and wave excitation energy to the transmission system. After arranging the first and second tuned mass dampers according to the optimal parameter combination obtained in this invention, the amplitudes of the above two resonance peaks are significantly reduced, and they broaden from sharp single peaks to gently sloping double peaks. Furthermore, no shift of response to other frequency bands is observed within the investigated frequency band. This indicates that the response energy of wind and wave input to the wind turbine and the axial displacement of the transmission system is effectively suppressed across the entire frequency range, consistent with the optimization intention of minimizing the H2 norm as the objective function in step S5. This verifies the effectiveness of the hybrid optimization strategy combining the pattern search algorithm and the simplex search algorithm.

[0059] 8. Numerical simulation verification of the optimal TMD parameter combination First, the optimal combination of tuned mass damper parameters obtained in step seven, namely the mass, stiffness, and damping parameters of the first and second tuned mass dampers, is backfilled and updated into the attributes of the fully coupled numerical model of the whole machine established in step one. A fully coupled time-domain simulation is then performed using environmental time-history excitations involving wind and wave coupling to extract the transmission system input loads, namely the six-degree-of-freedom hub loads and the six-degree-of-freedom nacelle motion.

[0060] Secondly, the extracted transmission system input loads and cabin motion data are used as external excitations and boundary conditions, and input into the rigid-flexible coupled multibody dynamics model of the transmission system established in step one. By performing time-domain dynamic simulation of the transmission system, the dynamic load time-history response of key internal components such as the main shaft, gears, and bearings is obtained.

[0061] Secondly, fatigue damage is assessed based on the component load response obtained from the transmission system simulation. The number of cycles of key bearings in the transmission system in each load range is obtained by dynamic equivalent load conversion, and the long-term cumulative fatigue damage degree of key components in the transmission system is calculated by combining Palmgren-Miner linear cumulative damage theory and wind field operating probability.

[0062] Finally, the comprehensive load reduction index output from the fully coupled numerical model of the whole machine, including the fatigue damage assessment value mentioned above, is compared and verified with the expected threshold of the engineering. If all load reduction performance indicators meet or exceed the target settings, the final tuned mass damper parameter configuration scheme is output; if the performance indicators do not meet the standards or there is a tuning deviation, the set of parameters is used as a new initial value and fed back to the pattern search algorithm in step six or the simplex search algorithm in step seven for secondary iterative optimization until the engineering design requirements are met, thus forming a complete technical process of parameter optimization, high-fidelity simulation verification and closed-loop feedback.

[0063] This step involves backfilling the optimized parameters into a high-fidelity model for time-domain simulation verification, and directly using the long-term cumulative fatigue damage of key components of the transmission system as the final judgment index. This ensures that the optimization results are not only optimal in the sense of frequency domain transfer function, but also have a real transmission system load reduction effect under actual wind and wave coupling conditions, providing a reliable parameter tuning scheme for the engineering application of deep-sea floating wind turbines.

[0064] Figure 4To compare the time-domain verification results of the parameter tuning effect of this invention, the method of obtaining the results is as follows: Following this step, the optimal parameter combination obtained through optimization is backfilled into the fully coupled numerical model of the whole machine and a fully coupled time-domain simulation is performed. The input load of the transmission system and the nacelle motion are extracted as boundary conditions and input into the rigid-flexible coupled multibody dynamics model of the transmission system. The dynamic load time history response of the upwind and downwind main bearings is solved. Then, the number of cyclic loadings in each load interval is statistically calculated through dynamic equivalent load conversion to obtain the dynamic equivalent load distribution of the main bearings. This result is then compared with the result obtained under the original configuration using the same procedure. Figure 4 (a) in the diagram represents the main bearing on the upwind side. Figure 4 In the diagram, (b) represents the leeward main bearing; the dashed line represents the original configuration, and the solid line represents the configuration after arranging the dual-tuned mass dampers with the optimal parameter combination. (Reference) Figure 4 It can be seen that after arranging the dual-tuned mass damper tuned by the present invention, the load distribution curves of the two main bearings shift as a whole towards the low load direction, the number of cycles in the high load range is significantly reduced, and the number of cycles in the low load range increases accordingly.

[0065] Because fatigue damage in rolling bearings is highly sensitive to load amplitude, reducing the number of cycles in the high-load range significantly reduces cumulative damage. Figure 4 Based on the dynamic equivalent load distribution shown, long-term cumulative fatigue damage was calculated using Palmgren-Miner linear cumulative damage theory and wind field operating probability. The results show that, compared to the original configuration, the optimal parameter combination tuned using the method of this invention reduces the equivalent cumulative fatigue damage of the upwind and downwind main bearings by 20.65% and 22.97%, respectively. Figure 3 and Figure 4 It can be seen that the present invention takes the H2 norm of the wind and wave excitation to the axial displacement of the transmission system as the tuning target, suppresses the energy transfer from the platform pitch mode and the first-order forward and backward bending mode of the tower to the transmission system in the frequency domain, and transforms it into a substantial reduction of the high load cycle of the main bearing under the full coupling condition in the time domain, thereby realizing the directional load reduction and fatigue life extension of the key components of the transmission system.

[0066] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A method for tuning parameters of a tuned mass damper for a floating wind turbine, characterized in that, Includes the following steps: S1. Construct a dynamic model of the entire floating wind turbine and its transmission system: Establish a fully coupled numerical model of the entire turbine by equating the transmission system with a single-degree-of-freedom spring-damped system, and establish a rigid-flexible coupled multibody dynamic model of the transmission system separately. S2. Construct a linear time-domain dynamic model of the floating wind turbine; S3. Using the dynamic response data output by the fully coupled numerical model of the whole machine, the unknown parameters in the linear time-domain dynamic model are identified to obtain the parameter-calibrated linear time-domain dynamic model. S4. Based on the linear time-domain dynamic model calibrated by parameters, a frequency-domain dynamic model including a tuned mass damper is constructed. The tuned mass damper includes a first tuned mass damper arranged in the cabin and a second tuned mass damper arranged on the floating platform. S5. Using the stiffness and damping parameters of the first and second tuned mass dampers as optimization variables, and minimizing the H2 norm of the axial displacement of the wind turbine and transmission system under wind and wave load input as the objective function, a parameter optimization problem is established. S6. Use a pattern search algorithm to perform global optimization on the parameter optimization problem and obtain the region where the global optimal solution is located; S7. The simplex search algorithm is used to perform local fine optimization in the region where the global optimal solution is located, so as to obtain the optimal parameter combination of the first tuned mass damper and the second tuned mass damper.

2. The method for tuning parameters of a tuned mass damper for a floating wind turbine according to claim 1, characterized in that, Also includes: S8. The optimal parameter combination is backfilled into the fully coupled numerical model of the whole machine for fully coupled time-domain simulation. The input load of the transmission system and the motion data of the cabin are extracted as boundary conditions and input into the rigid-flexible coupled multibody dynamics model of the transmission system to solve the load response of the internal key components. The fatigue damage degree of the key components of the transmission system is used as the verification index of the optimal parameter combination.

3. The method for tuning the parameters of a tuned mass damper for a floating wind turbine according to claim 1, characterized in that, The parameter identification in step S3 includes: under free decay test, with the goal of minimizing the root mean square error between the tower front-to-back deformation response and platform roll response output by the fully coupled numerical model of the whole machine and the tower front-to-back deformation response and platform roll response output by the linear time-domain dynamic model, the unknown parameters in the linear time-domain dynamic model are identified.

4. The method for tuning the parameters of a tuned mass damper for a floating wind turbine according to claim 1, characterized in that, The frequency domain dynamic model in step S4 is a five-degree-of-freedom frequency domain dynamic model, which is composed of the three-degree-of-freedom dynamic model of the entire floating wind turbine superimposed with the single-degree-of-freedom motion equations of the first tuned mass damper and the second tuned mass damper.

5. The method for tuning the parameters of a tuned mass damper for a floating wind turbine according to claim 1, characterized in that, The objective function in step S5 is the sum of the H2 norm of the wind load transfer function and the H2 norm of the wave load transfer function, expressed as: in, The transfer function for the axial displacement of the wind turbine and transmission system caused by wind load; The transfer function for the axial displacement of the wind turbine and transmission system caused by wave load.

6. The method for tuning the parameters of a tuned mass damper for a floating wind turbine according to claim 1, characterized in that, The global optimization of the pattern search algorithm in step S6 includes: generating an initial search starting point in the constraint space, setting the initial grid size and step size, and calculating the objective function value of the trial point in each exploration direction; if the objective function value decreases, the grid step size is increased and the trial point is used as the new iteration center point; otherwise, the grid step size is reduced; when the grid step size is lower than the preset switching threshold, the pattern search ends and the current optimal point is output.

7. The method for tuning parameters of a tuned mass damper for a floating wind turbine according to claim 6, characterized in that, The local fine optimization of the simplex search algorithm in step S7 includes: taking the current optimal point output in the pattern search stage as the first vertex of the initial simplex, constructing the initial simplex in the neighborhood of the current optimal point; calculating the objective function values ​​of all vertices and sorting them according to their quality; performing reflection, expansion, contraction, or edge shrinking operations on the worst point to generate candidate vertices; replacing the candidate vertex if its objective function value is better than the current worst point; iterating in a loop until the convergence criterion is met, and outputting the final global optimal point.

8. The method for tuning parameters of a tuned mass damper for a floating wind turbine according to claim 2, characterized in that, The evaluation of the fatigue damage degree of key components of the transmission system in step S8 includes: using dynamic equivalent load conversion to obtain the number of cycles of key bearings in each load range, and combining linear cumulative damage theory and wind field operating probability to calculate the long-term cumulative fatigue damage degree of key components of the transmission system; if the fatigue damage degree meets the preset engineering threshold, the optimal parameter combination is output; otherwise, the optimal parameter combination is used as a new initial point to return to step S6 for re-optimization.

9. The method for tuning parameters of a tuned mass damper for a floating wind turbine according to claim 1, characterized in that, Step S5 further includes: setting upper and lower bound constraints for the optimization variables based on the overall load-bearing capacity of the floating wind turbine and the actual layout space inside the nacelle and platform.

Images