A method and device for health monitoring of an offshore wind turbine tower and a storage medium

Abstract

The present invention relates to the technical field of wind power generation equipment monitoring, and specifically relates to a method and device for health monitoring of an offshore wind power tower barrel and a storage medium. The method includes: collecting environmental noise and recording it as a time-domain signal, segmenting the time-domain signal to form multiple segments of time-domain signals; preprocessing all the time-domain signals; performing passive inverse filtering on the preprocessed time-domain signals to obtain a time-domain impulse response matrix; analyzing the modal coupling characteristics of the tower barrel, drawing a three-dimensional wave velocity distribution map, and performing health monitoring on the tower barrel. The present invention uses environmental noise caused by wave impact as an excitation source, without manual intervention. For the collected environmental noise, by combining time-domain, frequency-domain, and spatial-domain features, a time-domain impulse response matrix is finally obtained, which is beneficial to analyzing the overall modal coupling characteristics of the tower barrel and drawing a high-precision three-dimensional wave velocity distribution map, thereby achieving sub-millimeter detection accuracy and reducing the false alarm rate.

Description

Technical Field

[0001] The present invention relates to the technical field of wind power generation equipment monitoring, and particularly relates to a method and device for health monitoring of an offshore wind power tower barrel and a storage medium. Background Art

[0002] Offshore wind power is an important part of renewable energy. However, the offshore environment is complex, and the impacts of tides and corrosion, offshore wind loads, etc. will all affect the operation of the wind power generation system. Therefore, it is necessary to conduct online monitoring of the offshore wind power system. As the support structure of the wind power generation system and also the structure directly in contact with seawater, the stability of the structure of the wind turbine tower barrel directly affects the normal operation of the wind power generation system.

[0003] In the prior art, for the health detection of the offshore wind power tower barrel structure, active guided wave technologies are mainly used, such as acoustic emission monitoring, ultrasonic monitoring, etc., or high-power telescope monitoring, thermal imaging monitoring, etc. This method is limited by artificial experience, equipment accuracy and environmental impacts, and the effect is not good in actual use, mainly reflected in the following aspects:

[0004] 1) The dependence on the excitation source restricts the long-term monitoring efficiency:

[0005] Active guided wave technologies generally rely on piezoelectric ceramic arrays to generate artificial excitation signals, and their energy conversion efficiency is low. Traditional solutions require the configuration of high-power power modules, which are difficult to meet the off-grid power supply requirements of offshore platforms.

[0006] 2) The low-density sensor layout limits the damage identification accuracy:

[0007] The low-density sensor layout results in insufficient spatial resolution and cannot detect microcracks. Fiber Bragg grating sensors can achieve high-density layout, but the prior art mostly uses single-axial strain monitoring.

[0008] 3) Single-mode analysis is easily interfered by environmental noise:

[0009] Traditional methods rely on a single guided wave mode for damage diagnosis, ignoring the multi-mode coupling effect, and the false alarm rate is high.

[0010] Therefore, it is necessary to provide a method and device for health monitoring of an offshore wind power tower barrel and a storage medium based on passive guided wave imaging. Summary of the Invention

[0011] In order to solve the above technical problems existing in the prior art, the present invention provides a method and device for health monitoring of an offshore wind power tower barrel and a storage medium.

[0012] To achieve the above object, the technical solution of the present invention is as follows:

[0013] In a first aspect, the present invention provides a method for health monitoring of an offshore wind turbine tower, including:

[0014] S1. Synchronously collect ambient noise using two or more sensors, record the ambient noise as a time-domain signal, and segment the time-domain signal to form multiple segments of time-domain signals;

[0015] S2. Preprocess all the time-domain signals;

[0016] S3. Perform passive inverse filtering on the preprocessed time-domain signals, specifically including:

[0017] S301. Perform Fourier transform on each segment of the preprocessed time-domain signal to obtain a frequency-domain matrix ;

[0018] S302. Perform singular value decomposition on each frequency in the frequency-domain matrix to obtain a singular value matrix;

[0019] S303. Regularize the singular value matrix;

[0020] S304. Reconstruct the frequency-domain impulse response matrix and obtain the time-domain impulse response matrix through inverse Fourier transform;

[0021] S4. Analyze the modal coupling characteristics of the tower based on the time-domain impulse response matrix.

[0022] Further, the duration of the collected ambient noise is 10 s, the sampling frequency is 2.5 MHz, and after being saved as a time-domain signal, it is segmented according to a duration of 0.1 s for each segment.

[0023] Furthermore, the number of sensors used in step S1 is 90. According to the position of the tower in seawater and the division of the tower into the fully submerged zone, tidal range zone, and splash zone, 30 sensors are set in each zone.

[0024] Furthermore, in step S2, preprocessing all the time-domain signals specifically includes: performing downsampling, filtering, and normalization processing on the time-domain signals in sequence.

[0025] Furthermore, the downsampling is specifically: the sampling frequency is 100 KHz, filtered using an anti-aliasing FIR filter, and downsampled with a downsampling factor of 10.

[0026] Furthermore, the filtering is specifically: filtered using a Butterworth band-pass filter, with a passband of 20 - 200 kHz and a stopband attenuation > 40 dB.

[0027] Furthermore, the normalization is specifically:

[0028]

[0029] Wherein:

[0030] ;

[0031]

[0032] represents the time-domain signal corresponding to each sensor after downsampling and filtering, i represents the sensor number; represents the normalized time-domain signal; is the average value of the time-domain signals of all sensors, represents the standard deviation of the time-domain signals of all sensors.

[0033] Furthermore, the frequency-domain matrix in step S301 is:

[0034]

[0035] Obtain the data matrix :

[0036]

[0037] Wherein, j is the imaginary unit, represents the k th frequency;

[0038] represents the frequency-domain matrix of the first sensor under the first time duration, represents the frequency-domain matrix of the first sensor under the 100th time duration; represents the frequency-domain matrix of the 90th sensor under the first time duration, represents the frequency-domain matrix of the 90th sensor under the 100th time duration.

[0039] Furthermore, in step S302, each frequency in the frequency-domain matrix is subjected to singular value decomposition to obtain a singular value matrix; specifically including:

[0040] Perform singular value decomposition through the following formula:

[0041]

[0042] Wherein, represents the transpose matrix of the frequency-domain matrix , represents the kth frequency;

[0043] represents the left singular vector, represents the transpose of the left singular vector, represents the singular value matrix.

[0044] Furthermore, in step S303, the singular value matrix is ​​regularized, specifically:

[0045] Setting Thresholds , such as the singular value matrix The singular values ​​in satisfy the following criteria:

[0046]

[0047] Among them, the singular value matrix is a matrix whose values ​​are all zero except for the diagonal elements. The diagonal elements are singular values, and the singular values ​​are arranged in descending order; is the first singular value, is the M+1th singular value, and from the first singular value onwards, until the M value is obtained when the judgment condition is met for the first time, and then M singular values ​​are obtained;

[0048] Singular values ​​that do not meet the determination condition are set to zero.

[0049] Furthermore, in step S304, the frequency domain impulse response matrix is ​​reconstructed by:

[0050]

[0051] In the above formula, represents the frequency domain impulse response matrix, is the M×M identity matrix, is the pseudo-inverse matrix that retains the first M singular values.

[0052] Furthermore, the time domain impulse response matrix is:

[0053]

[0054] in, represents the time domain impulse response matrix, represents the inverse Fourier transform.

[0055] Furthermore, the offshore wind turbine tower health monitoring method further includes:

[0056] S5. Guided wave imaging, specifically including:

[0057] S501, locate and quantify defects:

[0058] According to the frequency domain impulse response matrix and the time domain impulse response matrix, draw an image, which contains the dispersion curve, identify and extract the signal propagation characteristics; objective function Indicates the phase velocity of the area where defects are detected Relative phase velocity with respect to the defect-free region Square of the relative change in:

[0059]

[0060] : Wave number of the region where a defect has occurred, which is inversely proportional to the phase velocity ( , being the frequency after processing by the passive inverse filtering method of the received signal); ;

[0061] v 0: Wave number and phase velocity of the defect-free region; i being the piezoelectric sensor number;

[0062] S502. Based on the objective function Draw a guided wave imaging map.

[0063] In a second aspect, the present invention also provides an offshore wind turbine tower health monitoring device, comprising:

[0064] A sensor group, arranged on the inner wall of the tower, for collecting environmental noise;

[0065] A signal acquisition module, connected to the sensor group, for collecting the environmental noise received by the sensor group;

[0066] A receiving circuit, connected to the signal acquisition module, for receiving the environmental noise and transmitting the environmental noise signal to the acquisition circuit;

[0067] An acquisition circuit, connected to the receiving circuit, for transmitting the environmental noise signal to the processing center;

[0068] A processing center, connected to the acquisition circuit, for real-time processing and storage of data, and at the same time, based on the received environmental noise signal, monitoring the health status of the offshore wind turbine tower by using the above-mentioned offshore wind turbine tower health monitoring method.

[0069] Further, the sensor group includes 90 piezoelectric sensors. The tower is divided into three sections according to the fully submerged area, the tidal range area, and the splash zone, and 30 piezoelectric sensors are arranged on the inner wall of each section of the tower; and the 30 piezoelectric sensors on the inner wall of each section of the tower are arranged in an array along the circumferential direction of the tower.

[0070] Further, the piezoelectric sensors are adhered to the inner wall of the tower by epoxy resin adhesive.

[0071] In a third aspect, the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the above-mentioned method for health monitoring of an offshore wind turbine tower is implemented.

[0072] Compared with the prior art, the present invention has the following beneficial effects:

[0073] The method for health monitoring of an offshore wind turbine tower by passive guided wave imaging provided by the present invention uses environmental noise caused by wave impact as an excitation source. When the wave impacts the tower, an elastic wave is generated, and this signal will be received by the sensor. After signal processing to extract valuable high-frequency acoustic wave characteristics, relevant information about defects, corrosion, and cracks can be obtained. Without manual intervention, for the collected environmental noise, by combining time-domain, frequency-domain, and spatial-domain characteristics, a time-domain impulse response matrix is finally obtained, which is beneficial for analyzing the overall modal coupling characteristics of the tower and drawing a high-precision three-dimensional wave velocity distribution map, thereby achieving sub-millimeter detection accuracy and reducing the false alarm rate.

[0074] The present invention segments the tower, and a plurality of sensors are arranged on each segment. The outside of the tower is covered with high-risk corrosion areas (fully immersed area, tidal range area, atmospheric area), and a 3×30 sensor array is arranged inside the tower, so as to be able to achieve all-round real-time monitoring of the risk areas of the tower with high monitoring efficiency.

[0075] The present invention can monitor the corrosion rate, crack initiation, and propagation mechanism online in real time, and uses environmental noise caused by wave impact as an excitation source, with low power consumption and low cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 It is a flowchart of the method of the present invention.

[0077] Figure 2 It is a device architecture diagram of the present invention.

[0078] Figure 3 It is a schematic diagram of the tower segmentation of the present invention.

[0079] Figure 4 It is a schematic diagram of the installation position of the sensor group of the present invention.

[0080] Figure 5 It is a schematic diagram of the distribution of sensors on each layer of the present invention.

[0081] Reference numerals:

[0082] 1. Tower, 2. Piezoelectric sensor, 3. Piezoelectric sensor connection wire, 4. Signal acquisition module. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0083] The technical solution of the present invention will be clearly described below in conjunction with the accompanying drawings. Obviously, the described embodiments are not all embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0084] It should be noted that unless otherwise specifically stated, the relative arrangements of components and steps described in these embodiments and numerical expressions should not be construed as limiting the scope of the present invention.

[0085] The following description of exemplary embodiments is merely illustrative and in no way restricts the present invention and its application or use. Technologies, methods, and devices known to those of ordinary skill in the relevant fields may not be discussed in detail here, but when applicable, these technologies, methods, and devices should be regarded as part of this specification.

[0086] Embodiment 1

[0087] When an offshore wind turbine tower is subjected to stress, if it has defects, corrosion, and cracks, it will emit "transient elastic waves". This signal will be received by an acoustic emission sensor. By processing the signal to extract valuable high-frequency acoustic wave characteristics, relevant information about the occurrence of defects, corrosion, and cracks can be obtained. Install acoustic emission sensors on the offshore wind turbine tower. The damaged part of the tower emits elastic waves during the impact of sea waves and sea winds, which are captured by the sensors. Through data analysis and processing, the monitoring of the tower can be realized.

[0088] Therefore, this embodiment provides a method for health monitoring of an offshore wind turbine tower, as Figure 1 shown, including:

[0089] S1. Use more than two sensors to synchronously collect ambient noise, record the ambient noise as a time-domain signal, and segment the time-domain signal to form multiple segments of time-domain signals;

[0090] In this embodiment, 90 sensors are used to synchronously collect ambient noise. The 90 sensors are divided into 3 groups, with each group including 30 sensors. The 3 groups of sensors are respectively arranged in the tidal difference area, full immersion area, and atmospheric area of the tower. All sensors sample and record the ambient noise at a sampling frequency of 2.5 MHz for 10 seconds, and then perform post-processing to obtain the passive signals coupled by all sensors.

[0091] Save the continuously collected ambient noise (generated by the impact of wind and waves) as a time-domain signal , where, = 1, 2,..., 90. Divide the 10-second time-domain signal into 100 segments, each segment with a duration of 0.1 s to avoid transient interference. Construct a data matrix , the data matrix is a 90×100 matrix, and each column represents the time domain representation of a single noise segment:

[0092]

[0093] Self - constructed matrix , represents the transpose of the matrix, t represents time;

[0094] represents the time domain signal of the first sensor under the first time duration, represents the time domain signal of the first sensor under the second time duration, represents the time domain signal of the first sensor under the 100th time duration, represents the time domain signal of the second sensor under the first time duration, represents the time domain signal of the second sensor under the second time duration, represents the time domain signal of the second sensor under the 100th time duration, represents the time domain signal of the 90th sensor under the first time duration, represents the time domain signal of the 90th sensor under the second time duration, represents the time domain signal of the 90th sensor under the 100th time duration.

[0095] S2. Preprocess all time domain signals; specifically include: successively perform downsampling, filtering, and normalization on the time domain signals.

[0096] The downsampling: The original sampling rate of 2.5MHz is too high and is reduced to = 100KHz (retain the guided wave frequency band of 20–200kHz), and after using an anti - aliasing FIR filter, perform downsampling with a downsampling factor of 10.

[0097] Band - pass filtering: Use a Butterworth band - pass filter for filtering, with a passband of 20 - 200kHz and a stop - band attenuation > 40dB to filter out low - frequency vibrations and high - frequency electromagnetic noise.

[0098] Normalization: Remove the mean and normalize the signals for each channel:

[0099]

[0100] where:

[0101] ;

[0102]

[0103] Represents the time-domain signal corresponding to each sensor after downsampling and filtering. i Represents the sensor number. Represents the time-domain signal after normalization. Is the average value of the time-domain signals of all sensors. Represents the standard deviation of the time-domain signals of all sensors. Through normalization processing, the signal quality is improved and the algorithm stability is enhanced.

[0104] S3. Perform passive inverse filtering on the preprocessed time-domain signal, specifically including:

[0105] S301. Perform Fourier transform on each segment of the preprocessed time-domain signal to obtain a frequency-domain matrix :

[0106]

[0107] Obtain a data matrix :

[0108]

[0109] Where j Is the imaginary unit. Represents the k th frequency.

[0110] Represents the frequency-domain matrix of the first sensor under the first time duration, Represents the frequency-domain matrix of the first sensor under the 100th time duration; Represents the frequency-domain matrix of the 90th sensor under the first time duration, Represents the frequency-domain matrix of the 90th sensor under the 100th time duration.

[0111] S302. Perform singular value decomposition on each frequency in the frequency-domain matrix to obtain a singular value matrix; specifically including:

[0112] Perform singular value decomposition through the following formula:

[0113]

[0114] Where Represents the transpose matrix of the frequency-domain matrix ;

[0115] Represents the left singular vector, Represents the transpose of the left singular vector, Represents the singular value matrix.

[0116] S303, regularizing the singular value matrix; specifically:

[0117] Setting Thresholds , such as the singular value matrix The singular values ​​in satisfy the following criteria:

[0118]

[0119] Among them, the singular value matrix is a matrix whose values ​​are all zero except for the diagonal elements. The diagonal elements are singular values, and the singular values ​​are arranged in descending order; is the first singular value, is the M+1th singular value, and from the first singular value onwards, until the M value is obtained when the above judgment condition is met for the first time, and then M singular values ​​are obtained;

[0120] Singular values ​​that do not meet the judgment criteria are set to zero.

[0121] S304, reconstructing the frequency domain impulse response matrix, and obtaining the time domain impulse response matrix by inverse Fourier transform; the specific method is:

[0122]

[0123] In the above formula, represents the frequency domain impulse response matrix, is the M×M identity matrix, is the pseudo-inverse matrix that retains the first M singular values.

[0124] The time domain impulse response matrix is ​​obtained by inverse Fourier transform:

[0125]

[0126] in, represents the time domain impulse response matrix, represents the inverse Fourier transform.

[0127] S4. Analyze the modal coupling characteristics of the tower based on the time domain impulse response matrix.

[0128] First, perform intra-layer correlation, and use the above steps independently for each layer to generate three impulse response matrices , , , used to monitor the corrosion condition of each layer.

[0129] Then perform cross-layer association and combine the three-layer response matrices into:

[0130]

[0131] Analyze the overall modal coupling characteristics of the tower barrel.

[0132] S5, Guided wave imaging

[0133] S501, Locate and quantify defects:

[0134] The computer draws 3 graphs respectively from the data (one graph for the signal data of each layer), and 30 dispersion curves of guided waves corresponding to each layer are drawn in each graph, and the signal propagation characteristics (phase velocity, etc.) are identified and extracted.

[0135] Objective function represents the phase velocity of the area where a defect has occurred relative to the phase velocity of the defect-free area Relative change squared:

[0136]

[0137] : The wave number of the area where a defect has occurred, and the phase velocity is inversely proportional to ( , w is the frequency after the received signal is processed by the passive inverse filtering method); ;

[0138] : The wave number and phase velocity of the defect-free area; i is the piezoelectric sensor number.

[0139] Objective function The magnitude of measures the defect condition of the tower barrel: The larger it is, the more serious the defect of the tower barrel, thus locating and quantifying the defect.

[0140] S502, Image generation and visualization:

[0141] First, normalize the data to the range of [0, 1]:

[0142]

[0143] Maximum value , is the normalized value.

[0144] Then perform color mapping:

[0145] The RGB color mapping coordinates are (R, G, B);

[0146] Color of defect-free: : RGB color: dark blue (0, 0, 255).

[0147] Color of the defect and the intensity of the red channel is positively correlated with the normalized value: Red component: R = 255 × , RGB color: (255 × , 0, 0).

[0148] In summary, the objective function of receiving signals through the corresponding piezoelectric sensors changes, so as to locate and quantify the corrosion defect situation.

[0149] Embodiment 2

[0150] This embodiment provides a health monitoring device for an offshore wind turbine tower, including:

[0151] A sensor group, arranged on the inner wall of the tower, for collecting environmental noise.

[0152] A signal acquisition module, connected to the sensor group, for acquiring the environmental noise received by the sensor group;

[0153] A receiving circuit, connected to the signal acquisition module, for receiving environmental noise and delivering the environmental noise signal to the acquisition circuit;

[0154] An acquisition circuit, connected to the receiving circuit, for delivering the environmental noise signal to the processing center;

[0155] A processing center, connected to the acquisition circuit, for real-time processing and storage of data, and at the same time, based on the received environmental noise signal, monitoring the health status of the offshore wind turbine tower by using the above-mentioned health monitoring method for offshore wind turbine towers.

[0156] As Figures 2 to 5 shown, the sensor group includes 90 piezoelectric sensors 2. The tower 1 is divided into three sections according to the fully submerged area, the tidal range area and the splash zone, and 30 piezoelectric sensors 2 are arranged on the inner wall of each section of the tower 1; and the 30 piezoelectric sensors 2 on the inner wall of each section of the tower 1 are arranged in an array along the circumferential direction of the tower 1. The piezoelectric sensor 2 is a circular sensor with a radius of 18 mm and is adhered to the inner wall of the tower 1 by epoxy resin adhesive. All the piezoelectric sensors 2 are statistically connected to the piezoelectric sensor connection line 3 and then connected to the signal acquisition module 4.

[0157] Among them, the processing center includes a computer and an FPGA. The computer establishes communication with the FPGA through a wireless transmission protocol, and the FPGA is responsible for real-time processing and storage management of data as the core processor.

[0158] The signal acquisition module 4 is a 32-bit ADC with a sampling rate of 2.5 MHz, supporting synchronous acquisition of 90-channel signals; the FPGA chip realizes real-time signal preprocessing, including computing acceleration: using the FPGA parallelized passive inverse filtering algorithm and accelerating matrix operations; FPGA preprocessing (filtering, downsampling).

[0159] Embodiment 3

[0160] This embodiment provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned offshore wind turbine tower health monitoring method is realized.

[0161] The above specific embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the examples, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the scope of the technical solutions of the present invention, and they should all be covered by the scope of the claims of the present invention.

Claims

1. A method for health monitoring of an offshore wind turbine tower, characterized in that, Including: S1. Synchronously collect ambient noise using more than two sensors, record the ambient noise as a time-domain signal, segment the time-domain signal to form multiple segments of time-domain signals; The sensors used are arranged according to the position of the tower barrel in seawater and the division of the tower barrel into the fully submerged zone, tidal range zone, and splash zone, and sensors are set in each zone; S2. Preprocess all the time-domain signals; S3. Perform passive inverse filtering on the preprocessed time-domain signals, specifically including: S301. Perform Fourier transform on each preprocessed time-domain signal to obtain a frequency-domain matrix ; S302. Perform singular value decomposition on each frequency in the frequency-domain matrix to obtain a singular value matrix; S303. Regularize the singular value matrix; S304. Reconstruct the frequency-domain impulse response matrix and obtain the time-domain impulse response matrix through inverse Fourier transform; S4. Analyze the modal coupling characteristics of the tower barrel based on the time-domain impulse response matrix; S5. Guided wave imaging.

2. The method for health monitoring of an offshore wind turbine tower according to claim 1, wherein The duration of the collected ambient noise is 10 s, the sampling frequency is 2.5 MHz, and after being saved as a time-domain signal, it is segmented according to a duration of 0.1 s for each segment.

3. The method for health monitoring of an offshore wind turbine tower according to claim 2, wherein, The number of sensors used in step S1 is 90, and 30 sensors are set in each zone.

4. The method for health monitoring of an offshore wind turbine tower according to claim 3, characterized in that, In step S2, all the time-domain signals are preprocessed, specifically including: sequentially performing downsampling, filtering, and normalization processing on the time-domain signals.

5. The method for health monitoring of an offshore wind turbine tower according to claim 4, characterized in that, The specific downsampling is: the sampling frequency is 100 KHz, filtering is performed using an anti-aliasing FIR filter, and downsampling is performed with a downsampling factor of 10.

6. The method for health monitoring of an offshore wind turbine tower according to claim 4, wherein, The specific filtering is: filtering is performed using a Butterworth band-pass filter, the passband is 20 - 200 kHz, and the stopband attenuation > 40 dB.

7. The method for health monitoring of an offshore wind turbine tower according to claim 4, characterized in that The specific normalization is: Wherein: ； Represents the time-domain signal corresponding to each sensor after downsampling and filtering, i Represents the sensor number; Represents the normalized time-domain signal; Is the average value of the time-domain signals of all sensors, Represents the standard deviation of the time-domain signals of all sensors.

8. The method for health monitoring of an offshore wind turbine tower according to claim 7, characterized in that, The frequency domain matrix in step S301 is as follows: Obtain a data matrix : wherein, j is the imaginary unit, represents the k th frequency; Represents the frequency domain matrix of the first sensor at the first time period duration, Represents the frequency domain matrix of the first sensor at the 100th time period duration; Represents the frequency domain matrix of the 90th sensor at the first time period duration, Represents the frequency domain matrix of the 90th sensor at the 100th time period duration.

9. The method for health monitoring of an offshore wind turbine tower according to claim 3, wherein In step S302, singular value decomposition is performed on each frequency in the frequency-domain matrix to obtain a singular value matrix; specifically including: Performing singular value decomposition through the following formula: Among them, represents the transposed matrix of the frequency domain matrix , represents the k-th frequency; represents the left singular vector, represents the transpose of the left singular vector, represents the singular value matrix.

10. The method for health monitoring of an offshore wind turbine tower according to claim 9, wherein, In step S303, the singular value matrix is regularized, specifically: Set a threshold value , such as the singular value matrix in which the singular values satisfy the following determination conditions: Among them, the singular value matrix is a matrix whose values ​​are all zero except for the diagonal elements. The diagonal elements are singular values, and the singular values ​​are arranged in descending order; is the first singular value, is the M+1th singular value, and from the first singular value onwards, until the M value is obtained when the judgment condition is met for the first time, and then M singular values ​​are obtained; Singular values that do not meet the determination condition are set to zero.

11. The method for health monitoring of an offshore wind turbine tower according to claim 10, characterized in that, In step S304, the frequency-domain impulse response matrix is reconstructed, and the specific method is: In the above formula, represents the frequency-domain impulse response matrix, is an M×M identity matrix, is the pseudo-inverse matrix retaining the first M singular values.

12. The method for health monitoring of an offshore wind turbine tower according to claim 11, characterized in that, The time-domain impulse response matrix is: Among them, represents the time-domain impulse response matrix, represents the inverse Fourier transform.

13. The method for health monitoring of an offshore wind power tower barrel according to claim 12, wherein Step S5 specifically includes: S501. Locate and quantify defects: Based on the frequency-domain impulse response matrix and the time-domain impulse response matrix, an image is drawn, which includes dispersion curves, and the signal propagation characteristics are identified and extracted; the objective function represents the phase velocity of the area where defects are detected relative to the phase velocity of the defect-free area The square of the relative change of: : The wave number of the defective area is inversely proportional to the phase velocity ; , is the frequency after the received signal is processed by the passive inverse filtering method; ; : The wave number and phase velocity of the defect-free region; i is the piezoelectric sensor number; S502. Based on the objective function Draw a guided wave imaging map.

14. An offshore wind power tower health monitoring device, characterized in that, Including: A sensor group, arranged on the inner wall of the tower barrel, for collecting ambient noise; A signal acquisition module, connected to the sensor group, for acquiring the ambient noise received by the sensor group; A receiving circuit, connected to the signal acquisition module, for receiving the ambient noise and delivering the ambient noise signal to the acquisition circuit; An acquisition circuit, connected to the receiving circuit, for delivering the ambient noise signal to the processing center; A processing center, connected to the acquisition circuit, for real-time processing and storage of data, and at the same time, based on the received ambient noise signal, monitoring the health status of the offshore wind power tower barrel by using the method for health monitoring of an offshore wind power tower barrel according to any one of claims 1 - 13.

15. The health monitoring device for an offshore wind turbine tower according to claim 14, wherein The sensor group includes 90 piezoelectric sensors, the tower barrel is divided into three sections according to the fully submerged zone, tidal range zone, and splash zone, and 30 piezoelectric sensors are arranged on the inner wall of each section of the tower barrel; and the 30 piezoelectric sensors on the inner wall of each section of the tower barrel are arranged in an array along the circumferential direction of the tower barrel.

16. The health monitoring device for an offshore wind turbine tower according to claim 15, wherein The piezoelectric sensor is adhesively fixed on the inner wall of the tower barrel by using epoxy resin.

17. A computer-readable storage medium, characterized in that, A computer program is stored on the storage medium, and when the computer program is executed by a processor, the offshore wind turbine tower health monitoring method according to any one of claims 1-13 is implemented.

Images