Method for indirectly measuring hydrodynamic load of floating wind power platform with parameter self-adaption function

Abstract

The invention relates to a floating wind power platform hydrodynamic load indirect measurement method with a parameter adaptive function. The method comprises the following steps: (1) constructing a database based on working parameters, hydrodynamic loads and aerodynamic loads of a floating wind power platform; (2) training a set of neural network model taking the unit state and the environment state as input and the aerodynamic load as output based on a database; (3) when the relative error between the aerodynamic load output by the neural network and the aerodynamic load in the database is in a set range, the hydrodynamic load obtained in the step (1) in real time is a measurement result; and when the relative error between the aerodynamic load output by the neural network model and the aerodynamic load in the database exceeds a set range, adaptively adjusting the stiffness matrix parameter and obtaining the hydrodynamic load again. The rigidity parameter of the platform is adjusted in the middle and later stages of platform operation by using the independence of the neural network on the parameter, and then the hydrodynamic load of the platform can be accurately measured.

Description

Technical Field

[0001] The invention belongs to the technical field of load measurement of a floating wind power platform, and in particular relates to an indirect measurement method of a hydrodynamic load of a floating wind power platform with a parameter self-adaptation function. Background Art

[0002] After more than ten years of development, floating wind power technology has entered a critical stage of technical verification for large-scale commercialization and large-scale application, and faces many technical challenges. On the one hand, there are safety and reliability issues in complex environments: unlike traditional onshore wind turbines or fixed offshore wind turbines, floating wind turbines located in deep waters float on the sea surface and will face external input loads of different magnitudes and frequencies formed by wind, waves and currents, as well as endogenous loads generated by the movement of the power generation equipment in the platform system. On the other hand, there is the system cost issue under redundant design: in many current floating wind power projects, designers have adopted a higher safety factor in pursuit of excessive safety and stability, which makes the cost of the entire platform structure high, and it is inevitable to "correct" and "subtract" it. In fact, the design units and related companies are well aware of this, but in the current design, they have to stick to the rules due to the serious lack of support from load data under actual sea conditions. Whether it is system safety and stability regulation or structural cost reduction and optimization design, it is inseparable from the platform's load data feedback. Accurate load measurement data is conducive to grasping the balance between low-cost design and safety and stability of floating wind power platforms.

[0003] Floating wind power platforms are large-scale equipment, and the hydrodynamic loads they are subjected to are the result of the integration of hydrodynamic pressure. Such an overall load is difficult to measure directly using traditional force / torque sensors, and it is also impossible to install a large number of water pressure sensors on the platform surface at any cost. Currently, researchers mostly use numerical simulation methods to calculate the hydrodynamic loads borne by the platform, such as computational fluid dynamics based on the NS equation. This numerical simulation method can simulate the load of the platform under given sea conditions to a certain extent, but it cannot realize real-time calculation and feedback of the platform load under actual sea conditions. The extremely high requirements for computing resources limit its application in actual marine engineering.

[0004] The design life of floating wind power equipment is usually up to 20 years. During such a long operation process, some characteristic parameters inherent in the platform will inevitably change, such as the change of platform stiffness matrix parameters caused by the displacement of the platform anchor chain. Such parameter changes are difficult to be perceived by sensors, which will inevitably lead to inaccurate model calculations. Therefore, no matter what model is used, accurate evaluation of the characteristic parameters of the actual prototype is inevitable, which requires real-time adjustment of model parameters to adapt to changes in floating wind power system parameters. This process usually belongs to the first type of inverse problem of structural dynamics, namely system parameter identification. System parameter identification requires an additional set of system input and output (load-motion) data to complete the estimation of system parameters, and the lack of actual sea load data makes the parameter identification problem return to the load measurement technology itself. Therefore, there are already many neural network models to complete the load calculation of offshore wind turbines, but the accuracy of the neural network model depends largely on the training sample data, and the true value of the current floating wind power equipment load data is unknown, and there are very few actual operation data publicly released, which brings great challenges to the training of the model. Summary of the invention

[0005] In order to solve the current problems of inaccurate measurement of hydrodynamic loads on floating wind power platforms and the problem that platform stiffness matrix parameters change over time, the present invention provides an indirect load measurement method based on a load identification algorithm and a parameter adaptation method based on a neural network model.

[0006] The technical solution adopted by the present invention to solve its technical problem is:

[0007] An indirect measurement method for hydrodynamic load of a floating wind power platform with parameter self-adaptation function, the measurement method comprising the following steps:

[0008] (1) Construct a database based on the operating parameters, hydrodynamic loads, and aerodynamic loads of floating wind power platforms;

[0009] The working parameters include the six-degree-of-freedom motion state of the platform, the state of the unit, the state of the environment, and the constraint force of the platform;

[0010] The hydrodynamic load and the aerodynamic load are calculated based on the working parameters and the inherent mass matrix parameters, stiffness matrix parameters and damping matrix parameters of the platform and the inherent aerodynamic damping parameters of the impeller;

[0011] (2) training a neural network model based on the database, which takes the unit state and environmental state as input and the aerodynamic load as output; the neural network model is trained based on the database within a certain period of time after the floating wind power platform is put into operation;

[0012] (3) Compare the aerodynamic loads output by the neural network model with the aerodynamic loads in the database,

[0013] When the relative error between the aerodynamic load output by the neural network and the aerodynamic load in the database is within a set range, the hydrodynamic load obtained in real time in step (1) is the measurement result;

[0014] When the relative error between the aerodynamic load output by the neural network model and the aerodynamic load in the database exceeds a set range, the stiffness matrix parameters are adaptively adjusted and the hydrodynamic load is re-acquired.

[0015] Preferably, in step (1),

[0016] The six-degree-of-freedom motion state includes displacement, velocity and acceleration of six degrees of freedom;

[0017] The unit status includes impeller speed and blade pitch angle;

[0018] The environmental conditions include wind speed;

[0019] The platform restraining force is the mooring tension;

[0020] The specific calculation steps of the hydrodynamic load are: filtering the measured signal of the six-degree-of-freedom motion state, and calculating the hydrodynamic load of the floating wind power platform based on the load identification algorithm;

[0021] The specific calculation steps of the aerodynamic load are: based on the D'Alembert principle, the aerodynamic load of the floating wind power platform unit is calculated through the hydrodynamic load.

[0022] Preferably, in the specific calculation step of the hydrodynamic load, the six-degree-of-freedom motion state signal is band-pass filtered, the upper limit frequency of the filter is three times the frequency of the impeller rotor speed, and the lower limit frequency of the filter is the natural frequency of the platform;

[0023] The calculation formula of the load identification method is:

[0024]

[0025] Among them, F hydro represents the six-DOF hydrodynamic load on the platform, represents the Fourier transform operator, ω represents the circular frequency, t represents the time, x represents the six-degree-of-freedom displacement of the platform, C aero represents the aerodynamic damping matrix of the unit, and H(ω) represents the frequency response function of the platform system, which is expressed as follows:

[0026] H(ω)=(K-ω 2 M+jωC) -1 ;

[0027] Among them, M is the platform mass matrix, K is the platform stiffness matrix, C is the damping matrix of the platform, and j represents the imaginary part operator.

[0028] Preferably, in the specific calculation step of the aerodynamic load, the aerodynamic load derivation method based on the D'Alembert principle is given by the following formula:

[0029] F aero =-(F hydro +F moor +F inertia );

[0030] Among them, F aero Indicates the aerodynamic load, F moor Denotes the mooring load, F inertia represents the inertia force load, which can be calculated by the following formula:

[0031]

[0032] Preferably, the neural network model can be expressed as follows:

[0033] F aero =f NN (Ω,β,v);

[0034] Among them, Ω represents the impeller speed, β represents the blade pitch angle, and v represents the relative hub speed of the incoming flow;

[0035] The neural network model is trained based on a database of the floating wind power platform within half a year of being put into operation.

[0036] Preferably, in step (3), when the relative error between the aerodynamic load output by the neural network and the aerodynamic load calculated based on the D'Alembert principle exceeds 10%, the stiffness matrix parameters are adaptively adjusted.

[0037] Preferably, the process of adaptively adjusting the stiffness matrix parameters is as follows: measuring the six-degree-of-freedom motion state and unit state of the platform, calculating the aerodynamic load of the unit through a neural network model, low-pass filtering the motion state and aerodynamic load respectively, drawing a load-displacement curve, and calibrating the platform stiffness by solving the derivatives at different platform steady-state points, as shown below:

[0038]

[0039] Then, the stiffness matrix parameters in the load identification algorithm are updated to complete the subsequent hydrodynamic load measurement of the platform.

[0040] The beneficial effects of the present invention are as follows: the hydrodynamic load of the floating wind power platform is solved by collecting the platform motion state, unit state and environmental state which are easy to measure; then the aerodynamic load neural network model is trained based on the relatively accurate load data measured in the early stage of the platform operation, and then the platform stiffness parameters are adjusted in the middle and late stages of the platform operation by utilizing the independence of the neural network on the parameters, thereby ensuring the accurate measurement of the platform hydrodynamic load. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 This is a diagram of the hydrodynamic model of a floating wind power platform created in ANSYS-AQWA software.

[0042] Figure 2 It is a schematic diagram of the load and surge and pitch motions of a floating wind power system.

[0043] Figure 3 This is a CFD (fluid dynamics) simulation model diagram of a wind turbine in a floating wind power system.

[0044] Figure 4 This is the aerodynamic damping result diagram of the fan impeller under longitudinal and pitching motions.

[0045] Figure 5 It is the relationship diagram between the damping matrix parameters and frequency of the platform.

[0046] Figure 6 It is the data flow diagram of the hydrodynamic load measurement and neural network training in the initial stage of platform operation.

[0047] Figure 7 It is the structural diagram of the neural network model in this scheme.

[0048] Figure 8 It is the training flow chart for determining the optimal neural network model in this scheme.

[0049] Fig. 9 It is a data flow diagram for hydrodynamic load measurement and parameter adaptation in the middle and late stages of platform operation.

[0050] Fig.10 is the load-displacement curve, which is used to calibrate the stiffness coefficient.

[0051] Fig.11 It is a comparison diagram of the hydrodynamic load measurement and the measurement before and after parameter adaptation of the present invention. DETAILED DESCRIPTION

[0052] The present invention is further described below in conjunction with the accompanying drawings. This case focuses on the motion conditions of the sway and pitch degrees of freedom (the subscripts of the degrees of freedom are 1 and 5 of the six degrees of freedom, respectively). This implementation case is only one form of the present invention, and does not mean that the present invention is only applicable to the following situations. Any other implementation cases of the present invention are within the scope of protection of the present invention.

[0053] Reference Figures 1 to 11 , an indirect measurement method for hydrodynamic load of a floating wind power platform with parameter adaptive function, the measurement method comprising the following steps:

[0054] (1) Construct a database based on the operating parameters, hydrodynamic loads, and aerodynamic loads of floating wind power platforms;

[0055] The working parameters include the six-degree-of-freedom motion state of the platform, the state of the unit, the state of the environment, and the constraint force of the platform;

[0056] The hydrodynamic load and the aerodynamic load are calculated based on the working parameters and the inherent mass matrix parameters, stiffness matrix parameters and damping matrix parameters of the platform;

[0057] A large amount of measured unit status, environmental status and calculated aerodynamic load data is entered into the database;

[0058] (2) training a neural network model based on the database, taking the unit state and the environmental state as input and the aerodynamic load as output; the neural network model is trained based on the database within a certain period of time after the floating wind power platform is put into operation; in this embodiment, the neural network model is trained based on the database within half a year of the floating wind power platform being put into operation; the platform stiffness matrix parameters of the floating wind power platform remain unchanged or change little within half a year of being put into operation, and the load measurement data in this period of time is relatively accurate;

[0059] (3) Compare the aerodynamic loads output by the neural network model with the aerodynamic loads in the database,

[0060] When the relative error between the aerodynamic load output by the neural network and the aerodynamic load in the database is within a set range, the hydrodynamic load obtained in real time in step (1) is the measurement result;

[0061] When the relative error between the aerodynamic load output by the neural network model and the aerodynamic load in the database exceeds a set range, the stiffness matrix parameters are adaptively adjusted and the hydrodynamic load is re-acquired.

[0062] Preferably, in step (1),

[0063] The six-degree-of-freedom motion state includes displacement, velocity and acceleration of six degrees of freedom; an accelerometer, an inclinometer, a Beidou satellite positioning system, etc. are installed at a fixed position on the surface of the floating wind power platform to measure the six-degree-of-freedom displacement, velocity and acceleration of the platform, and then transform them into a geodetic coordinate system through coordinate transformation.

[0064] The unit status includes the impeller speed and the blade pitch angle; the data of the impeller speed and the blade pitch angle measured in this embodiment are shown in Table 1 below.

[0065] Table 1: Unit status data

[0066]

[0067] The environmental state includes the wind speed to which the floating wind power platform is subjected; the data of the wind speed measured in this embodiment are shown in Table 2 below.

[0068] Table 2 Environmental status data

[0069] Wind speed (m / s) 10.49057 9.829328 9.699986 10.01564 9.809997 9.949996 10.27548

[0070] The data of the six-degree-of-freedom motion state of the platform measured in this embodiment are shown in Table 3 below.

[0071] Table 3 Platform motion status data

[0072]

[0073]

[0074] The platform restraining force is the mooring tension; a tension sensor is installed on the catenary to measure the mooring line tension.

[0075] The specific calculation steps of the hydrodynamic load are: filtering the measured signal of the six-degree-of-freedom motion state, and calculating the hydrodynamic load of the floating wind power platform based on the load identification algorithm;

[0076] The specific calculation steps of the aerodynamic load are: based on the D'Alembert principle, the aerodynamic load of the floating wind power platform unit is calculated through the hydrodynamic load.

[0077] The mooring line load, hydrodynamic load, and aerodynamic load measured in this embodiment are shown in Table 4 below.

[0078] Table 4 Load data

[0079]

[0080] Preferably, in the specific calculation step of the hydrodynamic load, the six-degree-of-freedom motion state signal is band-pass filtered, the upper limit frequency of the filter is three times the frequency of the impeller rotor speed, and the lower limit frequency of the filter is the natural frequency of the platform; within this frequency domain, the platform motion response is excited by the platform hydrodynamic load and the aerodynamic damping load induced by the platform motion;

[0081] Then the calculation formula of the hydrodynamic load identification method is:

[0082]

[0083] Among them, F hydro represents the six-DOF hydrodynamic load on the platform, represents the Fourier transform operator, ω represents the circular frequency, t represents the time, and x represents the six-degree-of-freedom displacement of the platform; C aero It represents the aerodynamic damping matrix of the unit, which is determined by CFD simulation software; Figure 3 The figure shows a schematic diagram of the CFD model. The aerodynamic damping matrix is ​​determined as follows: In the flow field simulation software, a slight change in platform motion is given under different steady-state conditions, and the change in aerodynamic load is recorded, which can be written as:

[0084]

[0085] Figure 4 This is the aerodynamic damping result diagram of the fan impeller under longitudinal and pitch motions. b11 represents the longitudinal freedom degree, and b55 represents the pitch freedom degree. Aerodynamic damping is an impeller characteristic. The aerodynamic damping load is induced by the platform motion and is part of the aerodynamic load. Together with the hydrodynamic load, it stimulates the platform motion within the designed filter frequency band. Therefore, the influence of this part of the aerodynamic load needs to be eliminated in the hydrodynamic load identification formula.

[0086] H(ω) represents the acceleration frequency response function of the platform system, which is expressed as follows:

[0087] H(ω)=(K-ω 2 M+jωC) -1 ;

[0088] Where j is the imaginary operator; M is the platform mass matrix, K is the platform stiffness matrix, and C is the platform damping matrix, such as Figure 1 In the statics module of ANSYS software, a finite element model of the platform is created, and its mass matrix M is obtained. The mass matrix M is shown in Table 5 below;

[0089] Table 5. Mass matrix parameters

[0090] X Y Z RX RY R X 0.22268E+08 0.0000 0.0000 0.0000 -0.22499E+09 -0.20492E+06 Y 0.0000 0.22268E+08 0.0000 0.22499E+09 0.0000 -0.98187E+06 Z 0.0000 0.0000 0.22268E+08 0.20492E+06 0.98187E+06 0.0000 RX 0.0000 -0.22499E+09 -0.20492E+06 0.30321E+11 -0.17552E+07 0.47996E+07 RY 0.22499E+09 0.0000 -0.98187E+06 -0.17552E+07 0.30302E+11 70766. R 0.22499E+06 0.98187E+06 0.0000 0.47996E+07 70766. 0.20327E+11

[0091] C is the damping matrix parameter of the platform. The damping matrix parameter is the radiation damping matrix parameter, which is the inherent damping characteristic of the floating platform. Figure 5 The relationship between the radiation damping matrix parameters and the platform surge frequency is shown in Figure 2. R11 represents the surge degree of freedom, R55 represents the pitch degree of freedom, and their damping coefficients are derived through frequency domain analysis.

[0092] In the ANSYS-AQWA module, the hydrodynamic model and mooring model of the platform are created, and the load-displacement curve can be drawn through time domain analysis, such as Fig.10 As shown, it can be used to calibrate the stiffness coefficient and create the stiffness matrix K.

[0093] Preferably, Figure 2 The figure shows the force analysis diagram and motion diagram of the platform. In the specific calculation steps of the aerodynamic load, the aerodynamic load derivation method based on the D'Alembert principle is given by the following formula:

[0094] F aero =-(F hydro +F moor +F inertia );

[0095] Among them, F moor is the mooring load measured by the tension sensor, F hydro represents the hydrodynamic load indirectly measured in step (2), Indicates inertia force load; F inertia It can be calculated by the following formula:

[0096]

[0097] Preferably, Figure 6 It is a data flow diagram of platform hydrodynamic load measurement and neural network training. A database is built based on a large amount of measured unit status, environmental status and calculated aerodynamic load data in the early stage, and then a neural network model with unit status and environmental status as input and unit aerodynamic load as output is trained. The neural network model can be expressed as follows:

[0098] F aero =f NN (Ω,β,v);

[0099] Among them, Ω represents the impeller speed, β represents the blade pitch angle, and v represents the relative hub speed of the incoming flow; the training model uses a long-short cycle memory neural network, which is continuously iterated to determine the optimal sample characteristics (sampling period and time window), model parameters (number of network layers and units), and optimal training parameters (learning rate, solver, etc.); the trained neural network model does not depend on the system parameters and is separated from the database; the aerodynamic load calculated and output by the neural network model is relatively accurate and can be used as accurate reference data for subsequent calculations. Figure 7 and Figure 8 Shown is a schematic diagram of the long-short cycle neural network model and a training flow chart.

[0100] Preferably, in step (3), after the platform has been running for a long time, that is, in the middle and late stages of the operating cycle, the stiffness matrix parameters of the platform may change, and the hydrodynamic load and aerodynamic load measured by the load identification algorithm may be inaccurate; while the trained neural network does not depend on the system parameters, and the calculated output aerodynamic load is relatively accurate and can be used as a reference. Fig. 9 It is a data flow diagram for hydrodynamic load measurement and parameter adaptation in the middle and late stages of platform operation.

[0101] As shown in Table 6 below, when the relative error between the aerodynamic load output by the neural network and the aerodynamic load in the database is less than 10%, the calculation result can be directly stored in the database and output as a result;

[0102] Table 6. Data with error less than 10% without adjusting stiffness

[0103]

[0104] As shown in Table 7 below, when the relative error between the aerodynamic load output by the neural network and the aerodynamic load in the database exceeds 10%, it is determined that the error of the load measurement result is large, that is, the platform stiffness matrix parameters have changed and need to be adaptively adjusted.

[0105] Table 7. Data that require correction of stiffness when the error is greater than 10%

[0106]

[0107] Preferably, the process of adaptively adjusting the stiffness matrix parameters is as follows: measuring the six-degree-of-freedom motion state and the unit state of the platform, calculating the aerodynamic load of the unit through a neural network model, low-pass filtering the motion state and the aerodynamic load respectively, and drawing a load-displacement curve, wherein the ordinate of the load-displacement curve corresponds to the displacement of pitch and surge, and the abscissa corresponds to the aerodynamic load; calibrating the platform stiffness by solving the derivatives at different platform steady-state points, as shown below:

[0108]

[0109] like Fig.10 As shown, the stiffness matrix parameters in the load identification algorithm are updated to complete the subsequent hydrodynamic load measurement of the platform. At the same time, the aerodynamic load and hydrodynamic load obtained by recalculation are updated with the corresponding content in the database.

[0110] Fig.11The figure shows the comparison of the hydrodynamic load measurement results before and after parameter adaptation. It is easy to see from the results that after parameter adaptation, the platform hydrodynamic load calculation results are closer to the reference value, and the root mean square error is reduced from 16% to 9%, which shows the effectiveness of the indirect measurement method of the hydrodynamic load of the floating wind power platform with parameter adaptation function proposed by the present invention.

Claims

1. An indirect measurement method for hydrodynamic loads of a floating wind power platform with parameter adaptive function, characterized in that: The steps include: (1) Construct a database based on the operating parameters, hydrodynamic loads, and aerodynamic loads of floating wind power platforms; The working parameters include the six-degree-of-freedom motion state of the platform, the state of the unit, the state of the environment, and the constraint force of the platform; The hydrodynamic load and the aerodynamic load are calculated based on the working parameters and the inherent mass matrix parameters, stiffness matrix parameters and damping matrix parameters of the platform and the inherent aerodynamic damping parameters of the impeller; (2) training a neural network model based on the database, which takes the unit state and environmental state as input and the aerodynamic load as output; the neural network model is trained based on the database within a certain period of time after the floating wind power platform is put into operation; (3) Compare the aerodynamic loads output by the neural network model with the aerodynamic loads in the database, When the relative error between the aerodynamic load output by the neural network and the aerodynamic load in the database is within a set range, the hydrodynamic load obtained in real time in step (1) is the measurement result; When the relative error between the aerodynamic load output by the neural network model and the aerodynamic load in the database exceeds a set range, the stiffness matrix parameters are adaptively adjusted and the hydrodynamic load is re-acquired.

2. The indirect measurement method of hydrodynamic load of a floating wind power platform with parameter adaptive function according to claim 1, characterized in that: In the step (1), The six-degree-of-freedom motion state includes displacement, velocity and acceleration of six degrees of freedom; The unit status includes impeller speed and blade pitch angle; The environmental conditions include wind speed; The platform restraining force is the mooring tension; The specific calculation steps of the hydrodynamic load are: filtering the measured signal of the six-degree-of-freedom motion state, and calculating the hydrodynamic load of the floating wind power platform based on the load identification algorithm; The specific calculation steps of the aerodynamic load are: based on the D'Alembert principle, the aerodynamic load of the floating wind power platform unit is calculated through the hydrodynamic load.

3. The indirect measurement method of hydrodynamic load of a floating wind power platform with parameter adaptive function as claimed in claim 2, characterized in that: In the specific calculation step of the hydrodynamic load, the six-degree-of-freedom motion state signal is band-pass filtered, the upper limit frequency of the filter is three times the frequency of the impeller rotor speed, and the lower limit frequency of the filter is the natural frequency of the platform; The calculation formula of the load identification method is: Among them, F hydro represents the six-DOF hydrodynamic load on the platform, represents the Fourier transform operator, ω represents the circular frequency, t represents the time, x represents the six-degree-of-freedom displacement of the platform, C aero represents the aerodynamic damping matrix of the unit, and H(ω) represents the frequency response function of the platform system, which is expressed as follows: H(ω)=(K-ω 2 m+jωC) -1 ; Among them, M is the platform mass matrix, K is the platform stiffness matrix, C is the damping matrix of the platform, and j represents the imaginary part operator.

4. The indirect measurement method of hydrodynamic load of a floating wind power platform with parameter adaptive function as claimed in claim 3 is characterized by: In the specific calculation steps of the aerodynamic load, the aerodynamic load derivation method based on the D'Alembert principle is given by the following formula: F aero =-(F hydro +F moor +F inertia ); where F aero Indicates the aerodynamic load, F moor Denotes the mooring load, F inertia represents the inertia force load, which can be calculated by the following formula:

5. The indirect measurement method of hydrodynamic load of a floating wind power platform with parameter adaptive function according to claim 1, characterized in that: The neural network model can be expressed as follows: F aero =f NN (Ω,β,v); Wherein, Ω represents the impeller speed, β represents the blade pitch angle, and v represents the speed of the incoming flow relative to the hub; the neural network model is trained based on a database within half a year of the commissioning of the floating wind power platform.

6. The indirect measurement method of hydrodynamic load of a floating wind power platform with parameter adaptive function according to claim 1, characterized in that: In the step (3), when the relative error between the aerodynamic load output by the neural network and the aerodynamic load calculated based on the D'Alembert principle exceeds 10%, the stiffness matrix parameters are adaptively adjusted.

7. The indirect measurement method of hydrodynamic load of a floating wind power platform with parameter adaptive function as claimed in claim 6, characterized in that: The process of adaptive adjustment of the stiffness matrix parameters is as follows: measure the six-degree-of-freedom motion state and unit state of the platform, calculate the aerodynamic load of the unit through the neural network model, low-pass filter the motion state and aerodynamic load respectively, draw the load-displacement curve, and calibrate the platform stiffness by solving the derivatives at different platform steady-state points, as shown below: Then, the stiffness matrix parameters in the load identification algorithm are updated to complete the subsequent hydrodynamic load measurement of the platform.

Images