Automatic identification method for structural modal parameters of offshore wind turbine

Abstract

The invention provides an automatic identification method for modal parameters of an offshore wind turbine structure, and the method comprises the steps: carrying out the power spectrum density analysis of the actual measurement structure dynamic response of the offshore wind turbine structure, and recognizing the harmonic frequency through a long short-term memory (LSTM) network deep learning network; obtaining the dynamic response of the offshore wind turbine structure only excited by the environment by using a periodic sub-signal-Kalman filter; using a Monte Carlo covariance driven random subspace identification method to obtain a Monte Carlo stability graph, and further identifying the stability mode of the structure; and finally, obtaining modal parameters of the offshore wind turbine structure by using a density-based noisy application space clustering method. According to the invention, the challenge of harmonic excitation to modal parameter identification can be effectively solved, and the accuracy of the structural modal parameters is improved; and meanwhile, the structural health monitoring, state evaluation and damage identification capabilities of the offshore wind turbine are improved, so that the development of offshore wind energy is supported.

Description

Technical Field

[0001] This application belongs to the technical field of structural modal parameter identification, and relates to a modal parameter identification method, in particular to an automated identification method for the structural modal parameters of an offshore wind turbine. Background Technique

[0002] With the continuous transformation of the global energy structure, offshore wind turbines, as a clean and renewable energy source, are gradually attracting great attention from various countries. Due to the special environment where offshore wind turbines are located (including high wind speeds, large waves, marine corrosion, etc.), ensuring the safety of their structures and long-term stable operation is particularly important.

[0003] In this context, the accurate identification of the dynamic characteristics of offshore wind turbines, especially modal parameters (such as natural frequency, damping ratio, and vibration mode), has become a key link in ensuring their structural health, predicting fatigue damage, optimizing design, and improving operation efficiency. The modal identification of this structure can effectively help engineers timely understand the working state of the wind turbine and provide a scientific basis for subsequent maintenance and repair; at the same time, it provides data support for design optimization, ultimately improving the economic benefits and safety performance of the entire wind farm.

[0004] However, the identification of modal parameters of offshore wind turbines still faces many challenges. Traditional modal identification methods, such as experimental modal analysis and modal identification based on vibration data, although have achieved certain results in some applications, due to the complex working environment of offshore wind turbines, traditional methods usually rely on a large amount of experimental data, expensive equipment, and long-term operations, and the accuracy of their identification results is often affected by environmental noise and data quality.

[0005] To sum up, the currently adopted method for identifying the modal parameters of offshore wind turbines has problems such as low accuracy of modal parameter identification due to excessive manual intervention and inability to effectively guide engineering practice. Summary of the Invention

[0006] The purpose of this application is to provide an automated identification method for the structural modal parameters of an offshore wind turbine, which is used to solve the problems in the existing technology for identifying the modal parameters of offshore wind turbines, such as low accuracy of modal parameter identification due to excessive manual intervention and inability to effectively guide engineering practice.

[0007] In a first aspect, the present application provides a method for automatically identifying modal parameters of an offshore wind turbine structure, comprising the following steps: obtaining the dynamic response data of the offshore wind turbine structure in a target area; performing power spectral density analysis based on the dynamic response data of the offshore wind turbine structure to identify harmonic frequencies; performing filtering processing on the harmonic frequencies to obtain the dynamic response data of the offshore wind turbine structure under only environmental excitation; using the stochastic subspace algorithm to identify the dynamic response data of the offshore wind turbine structure under only environmental excitation to obtain a Monte Carlo stability diagram and the stable modal parameters of the offshore wind turbine structure; clustering based on the stable modal parameters of the offshore wind turbine structure to obtain the final modal parameters of the offshore wind turbine structure, so as to realize the automatic identification of the modal parameters of the offshore wind turbine structure.

[0008] In an implementation manner of the first aspect, performing power spectral density analysis based on the dynamic response data of the offshore wind turbine structure to identify harmonic frequencies includes: preprocessing the dynamic response data of the offshore wind turbine structure to obtain a time series segment of the dynamic response data; performing Fourier transform based on each time series segment and calculating the power spectral density to obtain a relationship curve between the power spectral density value and the frequency value of the dynamic response data; performing feature extraction on the relationship curve between the power spectral density value and the frequency value to obtain key features; the key features include: harmonic frequencies, spectral peak frequencies, and power spectral density amplitudes; constructing an LSTM model based on the key features; obtaining a training set and a test set of the time series segments of the dynamic response data and training and optimizing the LSTM model to obtain a trained LSTM model; identifying harmonic frequencies through the trained LSTM model.

[0009] In an implementation manner of the first aspect, the dynamic response data of the offshore wind turbine structure includes: the response of the offshore wind turbine structure under periodic excitation and the structural response data under only environmental excitation; performing filtering processing on the harmonic frequencies to obtain the dynamic response data of the offshore wind turbine structure under only environmental excitation includes: constructing a dynamic model of the offshore wind turbine structure under periodic excitation according to the harmonic frequencies; setting a periodic sub-signal Kalman filter based on the dynamic model; initializing the periodic sub-signal Kalman filter according to the harmonic frequencies and the dynamic response data of the offshore wind turbine structure; inputting the dynamic response data of the offshore wind turbine structure into the periodic sub-signal Kalman filter for filtering to obtain a periodic sub-signal corresponding to the harmonic frequencies; removing the periodic sub-signal from the dynamic response data of the offshore wind turbine structure to obtain the dynamic response data of the offshore wind turbine structure under only environmental excitation.

[0010] In an implementation of the first aspect, the random subspace algorithm is adopted to identify the dynamic response data of the offshore wind turbine structure under only environmental excitation, and the Monte Carlo stability diagram and the stable modes of the offshore wind turbine structure are obtained, including: determining the value range of control parameters based on the dynamic response data of the offshore wind turbine structure under only environmental excitation; the control parameters include: fundamental frequency, number of modes, time delay parameter, system order; using the Monte Carlo algorithm to simulate and generate several groups of random control parameters; using the covariance-driven random subspace identification algorithm to identify the random control parameters to obtain several groups of candidate modes; using the two-stage stability check method to identify based on several groups of the candidate modes to obtain stable mode parameters; and establishing a Monte Carlo stability diagram based on the stable mode parameters.

[0011] In an implementation of the first aspect, the value range of the control parameters includes: the value range of the time delay parameter and the value range of the system order; the calculation formula for the value range of the time delay parameter is:

[0012]

[0013] τ max = 2 × τ min

[0014] The calculation formula for the value range of the system order is:

[0015] N min = 2 × n m

[0016] N max = 2 × N min

[0017] Where, τ min represents the minimum value of the time delay parameter; τ max represents the maximum value of the time delay parameter; f s represents the sampling frequency value; f f represents the basic natural vibration frequency of the structure automatically obtained from the power spectral density diagram by using the scale space peak picking technique; N min represents the minimum value of the system order; N max represents the maximum value of the system order; n m represents the number of modes.

[0018] In an implementation of the first aspect, the two-stage stability check method includes: the first-stage preliminary stability check and the second-stage comprehensive stability check; the first-stage preliminary stability check is to compare each candidate mode with other candidate modes to screen out the modes that repeatedly appear under different parameter combinations and mark them as possible stable modes; the calculation formula used is:

[0019]

[0020] 1 - MAC(φ a , φ b ) ≤ 1%

[0021] Wherein, a and b respectively represent two different modes; f represents the natural vibration frequency; ξ represents the damping ratio; φ represents the vibration mode; Δf represents the relative difference in natural vibration frequency; Δξ represents the relative difference in damping ratio; MAC represents the modal assurance criterion between the vibration modes of two different modes.

[0022] In one implementation manner of the first aspect, the second - stage comprehensive stability check includes: classifying the parameters of the possible stable modes by using an algorithm to identify the stable - mode parameters; selecting the candidate modes with a stability index greater than 0.3 s as the stable results.

[0023] In one implementation manner of the first aspect, clustering is performed based on the stable - mode parameters of the off - shore wind turbine structure, and the final off - shore wind turbine structure modal parameters obtained include: obtaining the stable - mode parameter data set and the number of candidate modes based on the Monte Carlo stability diagram; calculating the value range of the minimum cluster size based on the number of candidate modes, and calculating the initial value of the minimum cluster size MinPts; making an optimal selection based on the initial value of MinPts to obtain the optimal setting under the current MinPts; performing clustering based on the optimal setting under the current MinPts to generate clustering clusters, and calculating the silhouette coefficient of the candidate modes in each clustering cluster; traversing all MinPts in turn to perform optimization parameter iteration to obtain the total silhouette coefficient; analyzing and optimizing the total silhouette coefficient to obtain the final off - shore wind turbine structure modal parameters.

[0024] In one implementation manner of the first aspect, calculating the value range of the minimum cluster size based on the number of candidate modes, and calculating the initial value of the minimum cluster size MinPts includes: calculating the initial value of MinPts of the minimum cluster size based on the number of candidate modes; for the current value of MinPts, calculating the distance between each data point in the stable - mode parameter data set obtained from the Monte Carlo stability diagram and its (MinPts - 1) - th nearest neighbor, and generating a K - distance graph; connecting the first point and the last point of the K - distance graph to form a first straight line; finding the point on the K - distance graph that is the farthest from the first straight line and perpendicular to the straight line as the elbow point; the distance value corresponding to the elbow point is the optimal setting value of the maximum clustering range.

[0025] In one implementation manner of the first aspect, the calculation formula of the silhouette coefficient is:

[0026]

[0027] Among them, Sil(p) represents the silhouette coefficient; b(p) represents the minimum value of the average distance from point p to other clusters; av(p) represents the average value of the distances from point p to other points within the same cluster.

[0028] As described above, the automatic identification method for the structural modal parameters of an offshore wind turbine in this application has the following beneficial effects:

[0029] (1) This application provides an automatic identification method for the structural modal parameters of an offshore wind turbine, which can effectively eliminate the influence of harmonic excitation on the identification of the modal parameters of the offshore wind turbine. Based on the use of power spectral density analysis in the actual measured dynamic response process of the offshore wind turbine structure, the harmonic frequency is automatically identified using an LSTM deep learning network; subsequently, based on the identified harmonic frequency, a periodic sub-signal Kalman filter is used to obtain the dynamic response of the offshore wind turbine structure under only environmental excitation; then, the Monte Carlo covariance-driven stochastic subspace identification method is used to automatically identify the modal parameters of the offshore wind turbine structure.

[0030] (2) The method provided in this application can, through the covariance-driven stochastic subspace method based on Monte Carlo, simultaneously consider the influence of system order parameters and time-delay parameters on the final modal identification result, thereby improving the accuracy of the identification result; in addition, it can also improve the automation degree and identification efficiency of the identification process by introducing a deep learning algorithm.

[0031] (3) In this application, the effectiveness and accuracy of this method are verified through numerical simulation and on-site measurement, which can provide an effective tool for the identification of the modal parameters of the offshore wind turbine structure under operating conditions, thereby improving the structural health monitoring, condition assessment, and damage identification capabilities of the offshore wind turbine. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It shows a schematic diagram of the hardware application scenario of the automatic identification method for the structural modal parameters of the offshore wind turbine described in this application in an embodiment.

[0033] Figure 2 It shows a schematic diagram of the process of the automatic identification method for the structural modal parameters of the offshore wind turbine described in this application in an embodiment.

[0034] Figure 3 It shows a schematic diagram of the S2 process in the automatic identification method for the structural modal parameters of the offshore wind turbine described in this application.

[0035] Figure 4 It shows a schematic diagram of the LSTM deep learning model in this application.

[0036] Figure 5 It shows a schematic diagram of the offshore wind turbine and its dynamic response time history diagram in this application.

[0037] Figure 6 It shows the schematic diagram of the S3 process in the automatic identification method of the structural modal parameters of the offshore wind turbine described in this application.

[0038] Figure 7 It shows the analysis diagram of the structural acceleration and power spectral density of the offshore wind turbine of this application.

[0039] Figure 8 It shows the schematic diagram of the S4 process in the automatic identification method of the structural modal parameters of the offshore wind turbine described in this application.

[0040] Figure 9 It shows the stability diagram of the structural response modal identification in the automatic identification method of the structural modal parameters of the offshore wind turbine described in this application.

[0041] Figure 10 It shows the schematic diagram of the S5 process in the automatic identification method of the structural modal parameters of the offshore wind turbine described in this application.

[0042] Figure 11 It shows the damping ratio - frequency relationship diagram of the stable candidate modes of the offshore wind turbine of this application.

[0043] Description of component labels

[0044] 11 Frequency analysis module

[0045] 12 Data processing module

[0046] 13 Spatial identification module

[0047] 14 Spatial clustering module Detailed implementation manners

[0048] The following uses specific specific examples to illustrate the implementation manners of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. This application can also be implemented or applied through other different specific implementation manners. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, without conflict, the following embodiments and the features in the embodiments can be combined with each other.

[0049] It should be noted that the diagrams provided in the following embodiments only illustrate the basic concept of this application in a schematic manner. Therefore, only the components related to this application are shown in the diagrams, rather than being drawn according to the number, shape, and size of the components in actual implementation. The types, quantities, and proportions of the components in actual implementation can be arbitrarily changed, and the component layout type may also be more complex.

[0050] The following embodiments of the present application provide an automated identification method for the structural modal parameters of an offshore wind turbine, which solves the problems in the existing modal parameter identification technology of offshore wind turbines, such as low accuracy of modal parameter identification due to excessive manual intervention and inability to effectively guide engineering practice.

[0051] As Figure 1 shown, the schematic diagram of the hardware application scenario of the automated identification method for the structural modal parameters of the offshore wind turbine described in the present application in an embodiment specifically includes: a frequency analysis module 11, a data processing module 12, a spatial identification module 13, and a spatial clustering module 14. Specifically, in the frequency analysis module 11, power spectral density analysis is performed on the dynamic response of the offshore wind turbine structure, and a harmonic frequency is identified using an LSTM (Long Short-Term Memory) deep learning network; through the data processing module 12, the identified harmonic frequency is used with a periodic sub-signal Kalman filter to obtain the dynamic response of the offshore wind turbine structure under only environmental excitation; in the spatial identification module 13, the Monte Carlo covariance-driven stochastic subspace identification method is applied to the structural dynamic response under only environmental excitation to obtain a Monte Carlo stability diagram, thereby identifying the stable modes of the structure; in the spatial clustering module 14, the density-based spatial clustering of applications with noise (DBSCAN) method is used to automatically obtain the modal parameters of the offshore wind turbine structure.

[0052] Next, the automated identification method for the structural modal parameters of the offshore wind turbine provided in the embodiments of the present application will be described in detail with reference to the accompanying drawings in the embodiments of the present application.

[0053] Please refer to Figure 2 , which shows the flow chart of the automated identification method for the structural modal parameters of the offshore wind turbine described in the present application in an embodiment. As Figure 2 shown, this embodiment provides an automated identification method for the structural modal parameters of an offshore wind turbine.

[0054] The automated identification method for the structural modal parameters of the offshore wind turbine specifically includes the following steps:

[0055] S1, obtain the dynamic response data of the offshore wind turbine structure in the target area.

[0056] In this embodiment, first, determine the specific location and scope of the offshore wind turbine for obtaining the dynamic response data. Then, according to the monitoring requirements, select appropriate sensors and monitoring devices (such as acceleration sensors, displacement sensors, force sensors, etc.) to collect the structural dynamic response data of the wind turbine.

[0057] Then, install the selected sensors and monitoring devices on the offshore wind turbine in the target area to ensure that the devices can collect data stably and accurately.

[0058] Finally, according to the monitoring requirements and data analysis requirements, set appropriate parameters such as sampling frequency and sampling duration; start the monitoring equipment installed on the wind turbine to collect the structural dynamic response data of the wind turbine; at the same time, record environmental parameters such as wind speed, wind direction, and waves in the target area for subsequent analysis.

[0059] S2. Perform power spectral density analysis based on the structural dynamic response data of the offshore wind turbine to identify harmonic frequencies. Please refer to Figure 3 , which shows the schematic diagram of the S2 process in the automatic identification method of the structural modal parameters of the offshore wind turbine described in this application. As Figure 3 shown, the S2 includes the following steps:

[0060] S21. Preprocess the structural dynamic response data of the offshore wind turbine to obtain time series segments of the dynamic response data;

[0061] S22. Perform Fourier transform based on each time series segment and calculate the power spectral density to obtain the relationship curve between the power spectral density value and the frequency value of the dynamic response data;

[0062] S23. Extract features from the relationship curve between the power spectral density value and the frequency value to obtain key features; the key features include: harmonic frequency, spectral peak frequency, and power spectral density amplitude;

[0063] S24. Build an LSTM model based on the key features;

[0064] S25. Obtain the training set and test set of the time series segments of the dynamic response data, and train and optimize the LSTM model to obtain the trained LSTM model;

[0065] S26. Identify harmonic frequencies through the trained LSTM model.

[0066] Please refer to Figure 4 and Figure 5 .

[0067] In this embodiment, power spectral density analysis is performed on the measured structural dynamic response of the offshore wind turbine, and an LSTM deep learning model is used to identify harmonic frequencies. In order to achieve efficient automatic identification, a deep learning framework is developed. According to this framework, the whole process is divided into six layers. A large number of samples are obtained based on on-site measurements and numerical simulation data. Based on training the LSTM deep learning network, the difference in the predicted harmonic frequencies is very small, and subsequent automatic structural analysis is ensured.

[0068] The developed deep learning framework consists of six layers: an input layer, a convolutional neural network layer, a long short-term memory network layer, an attention layer, a fully connected layer, and an output layer.

[0069] The input layer represents the power spectral density curve of the structural response under harmonic and random excitations, and the output layer represents the predicted harmonic frequencies.

[0070] Specifically, it is preferably to develop a deep learning framework based on LSTM. This deep learning framework can automatically identify and predict the harmonic frequencies in the dynamic response of the offshore wind turbine structure with a very small prediction difference to ensure the accuracy of subsequent automated structural analysis. That is: the deep learning framework includes: an input layer, a convolutional neural network layer, a long short-term memory network layer, an attention layer, a fully connected layer, and an output layer. Among them, the input layer receives the original data of the dynamic response of the offshore wind turbine structure, such as time series data of acceleration, displacement, etc.; the convolutional neural network layer uses methods such as power spectral density analysis to extract the features of the data, such as frequency components, power distribution, etc.; the long short-term memory network layer, as the core of the deep learning model, is used to capture the long-term dependencies in the time series data and identify the harmonic frequencies; the fully connected layer maps the output of the LSTM layer to the final predicted value, that is, the harmonic frequencies; finally, the predicted harmonic frequencies are output.

[0071] For example: it can be analyzed by the measured acceleration data of a certain offshore wind turbine located in the Yangjiang Wind Farm in China. That is: acceleration sensors are installed at three different height positions (such as 1#, 2#, 3#) on the offshore wind turbine structure to obtain the dynamic response and perform power spectral density analysis; using the LSTM deep learning model, a power spectral density analysis diagram is obtained, and the identified harmonic frequencies can prove the effectiveness of the deep learning model.

[0072] S3. Filter the harmonic frequencies to obtain the dynamic response data of the offshore wind turbine structure under only environmental excitations. Please refer to Figure 6 and Figure 7 , which are respectively shown as the schematic diagram of the S3 process in the method for automatically identifying the modal parameters of the offshore wind turbine structure described in this application and the analysis diagram of the acceleration and power spectral density of the offshore wind turbine structure of this application. As Figure 6 and Figure 7 shown, the S3 includes the following steps:

[0073] By determining the harmonic frequencies and using the periodic sub-signal - Kalman filter, the dynamic response of the offshore wind turbine structure under only environmental excitations is obtained. The dynamic response data of the offshore wind turbine structure includes: the response of the offshore wind turbine structure under periodic excitations and the response data of the structure under only environmental excitations.

[0074] S31. According to the harmonic frequencies, construct a dynamic model of the offshore wind turbine structure under periodic excitations;

[0075] S32. Based on the dynamic model, set the periodic sub-signal Kalman filter;

[0076] S33. Initialize the periodic sub-signal Kalman filter according to the harmonic frequency and the dynamic response data of the offshore wind turbine structure.

[0077] S34. Input the dynamic response data of the offshore wind turbine structure into the periodic sub-signal Kalman filter for filtering to obtain a periodic sub-signal corresponding to the harmonic frequency.

[0078] S35. Remove the periodic sub-signal from the dynamic response data of the offshore wind turbine structure to obtain the dynamic response data of the offshore wind turbine structure under only environmental excitation.

[0079] In this embodiment, the state of the system is estimated from the total structural response under combined periodic and random excitations through a Kalman filter; the harmonic states are identified based on the obtained harmonic frequencies, and periodic sub-signals are constructed from these harmonic states; by subtracting the identified periodic sub-signals from the total structural response, the structural dynamic response under random environmental excitation is obtained.

[0080] First, the harmonic frequencies obtained through power spectral density analysis are directly determined. By using a periodic sub-signal Kalman filter, the harmonic sub-signals are identified from the total structural dynamic response diagram and the power spectral density diagram. By subtracting the identified harmonic sub-signals from the structural response, the structural response under environmental excitation can be easily obtained. By initializing the state vector of the Kalman filter, its value can estimate the initial amplitude and phase of the signal through preliminary spectral analysis. Set the noise covariance matrix according to the noise characteristics of the system, and estimate the amplitude of the environmental noise through calibration data. At each moment, update the state vector according to the current observation value and prediction value, and gradually extract the amplitude and phase of the periodic signal. After being processed by the Kalman filter, the output periodic signal will contain the part of the structural response caused by environmental excitation, while removing the periodic components generated by the inherent characteristics of the structure.

[0081] Specifically, first establish a dynamic model of the offshore wind turbine structure under periodic excitation according to the extracted harmonic frequencies. This model should be able to reflect the vibration characteristics and response modes of the structure under harmonic excitation; design a periodic sub-signal Kalman filter according to the dynamic model of the offshore wind turbine structure and the harmonic frequency identification results; determine the state vector of the filter, including parameters such as the amplitude and phase of the harmonic signal; design the state transition matrix and the observation matrix to reflect the dynamic characteristics and observation characteristics of the dynamic response of the offshore wind turbine structure.

[0082] Then, initialize the state vector of the Kalman filter according to the harmonic frequency identification results and preliminary dynamic response data analysis; set the process noise covariance matrix and the measurement noise covariance matrix to reflect the uncertainties of the system dynamics and the measurement process.

[0083] Next, input the collected dynamic response data into the designed Kalman filter; based on the state transition matrix and the observation matrix, and combined with the statistical characteristics of the process noise and the measurement noise, the Kalman filter recursively updates the state vector.

[0084] Then, extract the periodic sub-signals corresponding to the harmonic frequencies through the output of the Kalman filter; these periodic sub-signals reflect the response characteristics of the offshore wind turbine structure under harmonic excitation.

[0085] Finally, subtract the extracted periodic sub-signals from the original dynamic response data; the obtained result is the dynamic response of the offshore wind turbine structure affected only by environmental excitation. Then verify the dynamic response data after removing the periodic sub-signals to ensure that it reflects the structural response caused only by environmental excitation; if the verification result is not satisfactory, the parameters of the Kalman filter can be adjusted or the harmonic frequencies can be re-identified to improve the extraction accuracy.

[0086] It should be noted that in-depth analysis of the dynamic response data affected only by environmental excitation can help understand the vibration characteristics, fatigue damage conditions, etc. of the offshore wind turbine structure under environmental excitation. The processed dynamic response data can be used in the health monitoring and early warning system of the offshore wind turbine structure to detect potential safety hazards in a timely manner and take corresponding measures. At the same time, based on the analysis results of the dynamic response data, the design of the offshore wind turbine structure can be optimized and improved to enhance its ability to resist environmental excitations such as wind and waves.

[0087] S4. Use the stochastic subspace algorithm to identify the dynamic response data of the offshore wind turbine structure under only environmental excitation, and obtain the Monte Carlo stability diagram and the stable modal parameters of the offshore wind turbine structure. Please refer to Figure 8 and Figure 9 , which are respectively shown as the flow schematic diagram of S4 in the method for automatic identification of modal parameters of an offshore wind turbine structure described in this application and the stable diagram of structural response modal identification in the method for automatic identification of modal parameters of an offshore wind turbine structure described in this application. As Figure 8 and Figure 9 shown, S4 includes the following steps:

[0088] S41. Based on the dynamic response data of the offshore wind turbine structure under only environmental excitation, determine the value range of the control parameters; the control parameters include: fundamental frequency, number of modes, time delay parameter, system order; the value range of the control parameters includes: the value range of the time delay parameter and the value range of the system order;

[0089] S42. Use the Monte Carlo algorithm to simulate and generate several groups of random control parameters;

[0090] S43. Use the covariance-driven stochastic subspace identification algorithm to identify the stochastic control parameters and obtain several groups of candidate modes.

[0091] S44. Use the two-stage stability check method to identify based on several groups of the candidate modes and obtain the stable mode parameters.

[0092] In this embodiment, the two-stage stability check method includes: the first-stage preliminary stability check and the second-stage comprehensive stability check.

[0093] The first-stage preliminary stability check is to compare each candidate mode with other candidate modes to screen out the modes that repeatedly appear under different parameter combinations and mark them as possible stable modes.

[0094] The second-stage comprehensive stability check includes: using an algorithm to classify the parameters of the possible stable modes, identifying the stable mode parameters; selecting the candidate modes with a stability index greater than 0.3 s as the stable results.

[0095] S45. Establish a Monte Carlo stability diagram based on the stable mode parameters.

[0096] In this embodiment, the Monte Carlo covariance-driven stochastic subspace identification method is applied to the structural response under ambient excitation to obtain a Monte Carlo stability diagram to identify the stable modes of the structure.

[0097] Specifically, use the Monte Carlo-based SSI-COV method; by considering the fundamental frequency f f and the number of modes n m the value ranges of the two control parameters, and then use Monte Carlo simulation to generate s = 100 groups of stochastic control parameters; use the generated control parameters and the structural modal responses under stochastic excitation for the covariance-driven stochastic subspace method; then, through the two-stage stability check, a Monte Carlo-based stability diagram can be established to identify the structural modes.

[0098] The calculation formula for the value range of the time-delay parameter τ is:

[0099]

[0100] τ max = 2×τ min (2)

[0101] The calculation formula for the value range of the system order Ο is:

[0102] N min = 2×n m (3)

[0103] N max= 2 × N min (4)

[0104] where τ min represents the minimum value of the time-delay parameter; τ max represents the maximum value of the time-delay parameter; f s represents the sampling frequency value; f f represents the fundamental natural vibration frequency of the structure automatically obtained from the power spectral density diagram using the scale-space peak picking technique; N min represents the minimum value of the system order; N max represents the maximum value of the system order; n m represents the number of modes.

[0105] Then, s = 100 sets of random numbers regarding the system order parameter and the time-delay parameter can be generated using Monte Carlo simulation, and s candidate modes are obtained in total by using the generated control parameters and candidate mode parameters for the covariance-driven stochastic subspace identification method.

[0106] For example, within the value range of the time-delay parameter τ and the system order Ο, s sets of random integers [τ, Ο] are generated using Monte Carlo simulation as control parameters. Then, for each set of random control parameters [τ, Ο], the covariance-driven stochastic subspace identification algorithm (SSI-COV) is used to analyze the structural dynamic response under ambient excitation only to obtain the candidate mode parameters corresponding to each set of control parameters, including natural vibration frequency, damping ratio, mode shape, etc.

[0107] Next, by adopting two-stage stability checking, a two-stage stable diagram screening method is proposed to eliminate the spurious modes to obtain a high-quality clear stable diagram.

[0108] The first stage is to compare and screen each candidate mode with other candidate modes from other groups using the following criteria, i.e.:

[0109] The calculation formula adopted is:

[0110]

[0111] 1 - MAC(φ a , φ b ) ≤ 1% (7)

[0112] where a and b represent two different modes; f represents the natural vibration frequency; ξ represents the damping ratio; φ represents the mode shape; Δf represents the relative difference in natural vibration frequency; Δξ represents the relative difference in damping ratio; MAC represents the modal assurance criterion between the mode shapes of two different modes.

[0113] The second stage: Further verify the preliminarily screened modes to ensure their stability.

[0114] In the two-stage stability diagram screening method, in the second stage, it is considered that only the candidate modes with a stability index greater than 0.3s = 30 are stable results.

[0115] Finally, spurious modes were effectively identified through a two-stage stability verification method. For the convenience of comparison, a traditional stability diagram was provided.

[0116] That is: Based on the results of the two-stage stability check, a Monte Carlo-based stability diagram is drawn. The x-axis of the stability diagram is the natural vibration frequency, and the y-axis is the serial number (or group number) of the random control parameters [τ, Ο] generated by the Monte Carlo simulation.

[0117] It can be seen from this that due to the interference of harmonic excitation, the basic structural modes cannot be accurately identified, and several pseudo-modes are misidentified as stable results, making it challenging to obtain accurate modal identification results. For the Monte Carlo-based stability diagram, due to the absence of harmonic excitation interference, the workload of structural modal parameter identification is greatly reduced.

[0118] It should be noted that in the data preprocessing stage, more appropriate filtering methods and parameters can be selected according to the characteristics of the data and the identification requirements; in the random subspace algorithm selection and training stage, different algorithms and parameter combinations can be tried to find the best identification scheme; when performing the two-stage stability check, the consistency of the test conditions should be ensured to accurately evaluate the stability of the modal parameters. The setting of the stability index and threshold should be adjusted according to the specific application scenarios and requirements. In the Monte Carlo simulation and stability diagram generation stage, the number of simulations and the value range of the control parameters can be adjusted as needed.

[0119] In addition, the generation of the Monte Carlo stability diagram and the determination of the stable modal of the offshore wind turbine structure may need to be realized with the help of professional software and tools, such as programming languages and data analysis software like MATLAB, Python, etc. These software and tools provide rich data processing, algorithm implementation, and visualization functions, which help to complete the identification task more efficiently.

[0120] S5. Cluster based on the stable modal parameters of the offshore wind turbine structure to obtain the final modal parameters of the offshore wind turbine structure, so as to realize the automatic identification of the modal parameters of the offshore wind turbine structure. Please refer to Figure 10 which shows the schematic flow diagram of S5 in the automatic identification method of the modal parameters of the offshore wind turbine structure described in this application. As Figure 4 shown, the S5 includes the following steps:

[0121] S51. Obtain the stable modal parameter dataset and the number of candidate modes based on the Monte Carlo stability diagram;

[0122] S52. Calculate the value range of the minimum cluster size based on the number of candidate modes, and calculate the initial value of the minimum cluster size MinPts; including: calculating the initial value of MinPts of the minimum cluster size based on the number of candidate modes; for the current value of MinPts, calculate the distance between each data point in the stable mode parameter dataset obtained from each Monte Carlo stability graph and its (MinPts - 1)-th nearest neighbor, and generate a K-distance graph; connect the first point and the last point of the K-distance graph to form a first straight line; find the point on the K-distance graph that is farthest from the first straight line and perpendicular to the straight line as the elbow point; the distance value corresponding to the elbow point is the optimal setting value of the maximum clustering range.

[0123] S53. Make an optimal selection based on the initial value of MinPts to obtain the optimal setting under the current MinPts.

[0124] S54. Perform clustering based on the optimal setting under the current MinPts to generate clustering clusters, and calculate the silhouette coefficient of the candidate modes in each clustering cluster.

[0125] The formula for the silhouette coefficient is as follows:

[0126]

[0127] where Sil(p) represents the silhouette coefficient; b(p) represents the minimum value of the average distance from point p to other clusters; av(p) represents the average value of the distances from point p to other points within the same cluster.

[0128] S55. Traverse all MinPts in sequence to perform optimization parameter iteration to obtain the total silhouette coefficient.

[0129] S56. Analyze and optimize the total silhouette coefficient to obtain the final structural modal parameters of the offshore wind turbine.

[0130] In this embodiment, the structural modal parameters of the offshore wind turbine are obtained by the density-based spatial clustering of applications with noise method. The method of combining the silhouette coefficient and the elbow rule to automatically and optimally select the two control parameters required for the DBSCAN density clustering method is used for processing.

[0131] Combined with the number of stable candidate modes obtained in the previous steps, according to the formula N min = 5ln(P); N max = 10ln(P), where P is the number of modes in the quasi-Monte Carlo-based clear stability graph. Calculate the possible value range of the optimal value of the minimum cluster size MinPts [N min , N maxCalculate the silhouette coefficient of the candidate modes for each clustering result according to formula (8). When drawing the K-distance graph, the value of K is MinPts - 1. A straight line is formed by connecting the first point and the last point of the K-distance graph. The point on the K-distance graph that is farthest from the straight line and perpendicular to the straight line is defined as the elbow of the K-distance graph, and its value is used as the optimal setting of the maximum clustering range ε.

[0132] This method is mainly divided into the following five steps:

[0133] (1) Start with MinPts = N min and at the same time, the corresponding value of ε can be optimally set by drawing the K-distance graph and finding the value corresponding to the elbow position of the K-distance graph;

[0134] (2) Execute the density-based clustering method to obtain the clustering result;

[0135] (3) Calculate the silhouette coefficient of each candidate mode, multiply the silhouette coefficient of the candidate mode classified as the outlier class by -1, and finally add up the silhouette coefficients of all candidate modes as the evaluation index of the clustering quality of this time;

[0136] (4) Gradually increase the value of MinPts and iterate steps (1), (2), and (3) until MinPts is N max and record the results of the sum of the silhouette coefficients of each candidate mode in each iteration;

[0137] (5) Select the optimal values of the two DBSCAN control parameters according to the maximum value in the sum of the silhouette coefficients recorded each time: the size of the smallest cluster MinPts and the maximum clustering range ε.

[0138] As can be seen from the above, each point in the clustering has at least MinPts in its neighborhood within the given radius ε, and the points that do not meet this requirement are regarded as outliers and removed. Using the stability diagram of Monte Carlo, the DBSCAN clustering method is used to cluster the stable modal parameters, remove the abnormal noise points, obtain the true modes, and finally obtain the modal parameters by taking the average.

[0139] Combined with the data in Table 1, as Figure 11 shown, it can be seen that the estimated modal parameters obtained by using the above method are compared and verified with the modal parameters obtained by the reference method HM-SSI.

[0140] Table 1 Statistical table of the automatic recognition results of offshore wind turbines

[0141]

[0142] The comparison of the above content shows that there is good consistency between the two groups of results, verifying the effectiveness of the method provided by this application.

[0143] The automatic identification method for modal parameters of an offshore wind turbine structure provided by this application can effectively eliminate the influence of harmonic excitation on the identification of modal parameters of the offshore wind turbine. Based on the use of power spectral density analysis during the measured dynamic response process of the offshore wind turbine structure, the harmonic frequency is automatically identified using an LSTM deep learning network; subsequently, based on the identified harmonic frequency, a periodic sub-signal Kalman filter is used to obtain the dynamic response of the offshore wind turbine structure under only environmental excitation; then, the Monte Carlo covariance-driven stochastic subspace identification method is used to automatically identify the modal parameters of the offshore wind turbine structure. This application can, through the covariance-driven stochastic subspace method based on Monte Carlo, simultaneously consider the influence of system order parameters and time-delay parameters on the final modal identification result, thereby improving the accuracy of the identification result; in addition, it can also improve the automation level and identification efficiency of the identification process by introducing deep learning algorithms. At the same time, this application verifies the effectiveness and accuracy of this method through numerical simulation and on-site measurement, and can provide an effective tool for the identification of modal parameters of the offshore wind turbine structure under operating conditions, thereby improving the structural health monitoring, condition assessment, and damage identification capabilities of the offshore wind turbine, effectively overcoming various disadvantages in the prior art and having high industrial utilization value.

[0144] The protection scope of the automatic identification method for modal parameters of an offshore wind turbine structure described in the embodiments of this application is not limited to the execution order of the steps listed in this embodiment. Any solution achieved by adding or subtracting steps of the prior art and replacing steps according to the principle of this application is included in the protection scope of this application.

[0145] In summary, an automatic identification method for modal parameters of an offshore wind turbine structure provided by this application has the following beneficial effects:

[0146] An automatic identification method for modal parameters of an offshore wind turbine structure provided by this application can effectively solve the challenges brought by harmonic excitation to modal parameter identification and improve the accuracy of structural modal parameters; it is expected to improve the structural health monitoring, condition assessment, and damage identification capabilities of the offshore wind turbine, and thus support the development of offshore wind energy. At the same time, this invention effectively overcomes various disadvantages in the prior art and has high industrial utilization value.

[0147] The above embodiments are only illustrative of the principles and effects of this application, and are not used to limit this application. Any person familiar with this technology can modify or change the above embodiments without departing from the spirit and scope of this application. Therefore, all equivalent modifications or changes completed by those with ordinary knowledge in the technical field without departing from the spirit and technical ideas disclosed by this application should still be covered by the claims of this application.

Claims

1. An automated identification method for modal parameters of an offshore wind turbine structure, characterized in that, It includes the following steps: Obtain the dynamic response data of the offshore wind turbine structure within the target area; Conduct power spectral density analysis based on the dynamic response data of the offshore wind turbine structure to identify harmonic frequencies; Perform filtering processing on the harmonic frequencies to obtain the dynamic response data of the offshore wind turbine structure under only environmental excitation; Use the stochastic subspace algorithm to identify the dynamic response data of the offshore wind turbine structure under only environmental excitation to obtain the Monte Carlo stability diagram and the stable modal parameters of the offshore wind turbine structure; Cluster based on the stable modal parameters of the offshore wind turbine structure to obtain the final modal parameters of the offshore wind turbine structure, so as to realize the automatic identification of the modal parameters of the offshore wind turbine structure.

2. The automated identification method for the structural modal parameters of an offshore wind turbine according to claim 1, wherein Conducting power spectral density analysis based on the dynamic response data of the offshore wind turbine structure to identify harmonic frequencies includes: Preprocess the dynamic response data of the offshore wind turbine structure to obtain time series segments of the dynamic response data; Conduct Fourier transform based on each time series segment and calculate the power spectral density to obtain the relationship curve between the power spectral density value and the frequency value of the dynamic response data; Extract features from the relationship curve between the power spectral density value and the frequency value to obtain key features; the key features include: harmonic frequency, spectral peak frequency, and power spectral density amplitude; Construct an LSTM model based on the key features; Obtain the training set and test set of the time series segments of the dynamic response data, and train and optimize the LSTM model to obtain the trained LSTM model; Identify harmonic frequencies through the trained LSTM model.

3. The automated identification method for the structural modal parameters of an offshore wind turbine according to claim 1, wherein The dynamic response data of the offshore wind turbine structure includes: the response of the offshore wind turbine structure under periodic excitation and the structural response data under only environmental excitation; performing filtering processing on the harmonic frequencies to obtain the dynamic response data of the offshore wind turbine structure under only environmental excitation includes: Construct a dynamic model of the offshore wind turbine structure under periodic excitation according to the harmonic frequencies; Set a periodic sub-signal Kalman filter based on the dynamic model; Initialize the periodic sub-signal Kalman filter according to the harmonic frequencies and the dynamic response data of the offshore wind turbine structure; Input the dynamic response data of the offshore wind turbine structure into the periodic sub-signal Kalman filter for filtering to obtain a periodic sub-signal corresponding to the harmonic frequencies; Remove the periodic sub-signal from the dynamic response data of the offshore wind turbine structure to obtain the dynamic response data of the offshore wind turbine structure under only environmental excitation.

4. The automated identification method for structural modal parameters of an offshore wind turbine according to claim 1, characterized in that Using the stochastic subspace algorithm to identify the dynamic response data of the offshore wind turbine structure under only environmental excitation to obtain the Monte Carlo stability diagram and the stable modes of the offshore wind turbine structure includes: Determine the value range of control parameters based on the dynamic response data of the offshore wind turbine structure under only environmental excitation; the control parameters include: fundamental frequency, number of modes, time delay parameter, system order; Use the Monte Carlo algorithm to simulate and generate several groups of random control parameters; Use the covariance-driven stochastic subspace identification algorithm to identify the random control parameters to obtain several groups of candidate modes; Use the two-stage stability check method to identify based on several groups of the candidate modes to obtain stable modal parameters; Based on the stable modal parameters, a Monte Carlo stability diagram is established.

5. The automated identification method for the structural modal parameters of an offshore wind turbine according to claim 4, characterized in that The value range of the control parameters includes: the value range of the time-delay parameter and the value range of the system order. The calculation formula for the value range of the time-delay parameter is: τ max = 2 × τ min The calculation formula for the value range of the system order is: N min = 2 × n m N max = 2 × N min Among them, τ min represents the minimum value of the time-delay parameter; τ max represents the maximum value of the time-delay parameter; f s represents the sampling frequency value; f f represents the basic natural vibration frequency of the structure automatically obtained from the power spectral density diagram by using the scale-space peak picking technique; N min represents the minimum value of the system order; N max represents the maximum value of the system order; n m represents the number of modes.

6. The automated identification method for structural modal parameters of an offshore wind turbine according to claim 4, characterized in that, The two-stage stability check method includes: the first-stage preliminary stability check and the second-stage comprehensive stability check. In the first-stage preliminary stability check, each candidate mode is compared with other candidate modes to screen out the modes that repeatedly appear under different parameter combinations and mark them as possible stable modes. The calculation formula adopted is: 1-MAC(φ a ,φ b ) ≤ 1% Among them, a and b respectively represent two different modes; f represents the natural vibration frequency; ξ represents the damping ratio; φ represents the vibration mode; Δf represents the relative difference in natural vibration frequency; Δξ represents the relative difference in damping ratio; MAC represents the modal assurance criterion between the vibration modes of two different modes.

7. The automated identification method for structural modal parameters of an offshore wind turbine according to claim 6, wherein The second-stage comprehensive stability check includes: Using an algorithm to classify the parameters of the possible stable modes and identify the stable modal parameters. Select the candidate modes with a stability index greater than 0.3 s as the stable results.

8. The automated identification method for structural modal parameters of an offshore wind turbine according to claim 1, characterized in that Based on the stable modal parameters of the offshore wind turbine structure, clustering is performed to obtain the final offshore wind turbine structure modal parameters, including: Based on the Monte Carlo stability diagram, a stable modal parameter data set and the number of candidate modes are obtained. Based on the number of candidate modes, the value range of the minimum cluster size is calculated, and the initial value of the calculated minimum cluster size MinPts is calculated. Based on the initial value of MinPts, an optimal selection is made to obtain the optimal setting under the current MinPts. Based on the optimal setting under the current MinPts, clustering is performed to generate clustering clusters, and the silhouette coefficient of the candidate modes in each clustering cluster is calculated. All MinPts are traversed in sequence for optimized parameter iteration to obtain the total silhouette coefficient. The total silhouette coefficient is analyzed and optimized to obtain the final offshore wind turbine structure modal parameters.

9. The automated identification method for the structural modal parameters of an offshore wind turbine according to claim 8, wherein Based on the number of candidate modes, calculating the value range of the minimum cluster size and calculating the initial value of the calculated minimum cluster size MinPts includes: Calculating the initial value of MinPts of the minimum cluster size based on the number of candidate modes. For the current value of MinPts, calculate the distance between each data point in the stable modal parameter data set obtained from each Monte Carlo stability diagram and its (MinPts - 1)th nearest neighbor, and generate a K-distance diagram. Connect the first point and the last point of the K-distance diagram to form the first straight line. On the K-distance diagram, find the point that is farthest from the first straight line and perpendicular to the straight line as the elbow point; the distance value corresponding to the elbow point is the optimal setting value of the maximum clustering range.

10. The method for automatically identifying the modal parameters of an offshore wind turbine structure according to claim 8, wherein The calculation formula for the silhouette coefficient is: Among them, Sil(p) represents the silhouette coefficient; b(p) represents the minimum value of the average distance from point p to other clusters; av(p) represents the average value of the distances from point p to other points within the same cluster.

Images