Wind power tower modal analysis method and system based on FMCW four-dimensional laser radar dual-stage scanning

Abstract

The invention discloses a wind power tower modal analysis method and system based on FMCW four-dimensional laser radar dual-stage scanning. According to the method, the unique capability of an FMCW laser radar for directly measuring the radial speed of a target based on the Doppler frequency shift principle is utilized, and three-dimensional space coordinates and speed information are synchronously obtained to form a four-dimensional point cloud. The method adopts a double-stage scanning strategy: in the first stage, the overall geometric morphology of the tower drum is quickly obtained through spiral scanning with an oscillation compensation item; and in the second stage, self-adaptive vertical line reciprocating scanning is performed based on the geometric model to obtain high-density time sequence speed data. A time sequence speed matrix is constructed through a weighted median algorithm, and modal parameters are calculated and extracted in combination with frequency domain analysis and fusion damping ratio. The contradiction between space coverage and time resolution in the prior art is solved, and compared with a traditional method, the scanning efficiency is improved by more than 80%, and the damping ratio recognition precision is improved by more than 50%.

Description

Technical Field

[0001] This invention relates to the field of structural health monitoring technology, specifically to a non-contact modal analysis method and system for wind turbine towers based on FMCW (Frequency Modulated Continuous Wave) four-dimensional lidar. Background Technology

[0002] Wind turbine towers are typical flexible, tall structures that withstand complex alternating loads such as wind and wave loads over long periods. Their structural health directly affects the safe operation of wind turbines. Modal analysis, by identifying dynamic characteristic parameters such as the structure's natural frequencies, damping ratios, and mode shapes, provides crucial information for structural safety assessment and fault diagnosis, and is a core technical means for structural health monitoring.

[0003] Existing modal analysis techniques for wind turbine towers mainly include the following categories: The first type is based on contact sensors, mainly using accelerometers. This method requires the installation of multiple sensors at different heights on the tower, which has the following problems: (1) installation is difficult and high-altitude operations are risky, especially in the harsh environment of offshore wind power tower installation; (2) the number of measuring points is limited by wiring and cost, making it difficult to obtain complete modal vibration information; (3) sensors and cables are easily damaged by long-term exposure to harsh environments, resulting in high maintenance costs.

[0004] The second category is non-contact measurement methods based on traditional lidar. Pulse lidar and phase lidar can only directly measure the three-dimensional spatial coordinates (x, y, z) of the target. Obtaining velocity information requires continuous position measurement and differential calculation, i.e. This indirect method has the following technical drawbacks: (1) The noise amplification effect is introduced by differential calculation. The random noise of displacement measurement is amplified after differential calculation. In particular, when Δt is small, the velocity noise increases significantly, which seriously affects the identification accuracy of modal parameters.

[0005] (2) The contradiction between time resolution and accuracy. To reduce differential noise, Δt needs to be increased, but this will reduce the time resolution of velocity and make it impossible to accurately capture high-frequency vibration components.

[0006] (3) The contradiction between spatial coverage and temporal resolution. Obtaining the complete tower geometry requires a slow full-coverage scan (usually taking more than 300 seconds), but modal analysis requires a high sampling frequency to capture rapid vibrations. Existing single scanning modes cannot simultaneously meet these two contradictory requirements.

[0007] FMCW lidar is a novel type of lidar based on the principle of coherent detection. Its core feature is the emission of a continuous laser signal whose frequency changes linearly with time. When the laser illuminates a moving target, the reflected echo experiences a Doppler frequency shift relative to the emitted signal. By analyzing the mixing result of the emitted and echo signals, FMCW lidar can simultaneously separate range information (from the frequency difference caused by the optical path difference) and velocity information (from the Doppler frequency shift) from the beat frequency signal. This means that FMCW lidar can synchronously and directly output four-dimensional data for each measurement point: three-dimensional spatial coordinates (x, y, z) and radial velocity (v), without needing to indirectly obtain the velocity through differential calculations.

[0008] However, there is currently a lack of modal analysis methods for wind turbine towers that optimize the design of four-dimensional measurement characteristics of FMCW lidar. Existing technologies have failed to fully leverage the unique advantages of FMCW lidar in direct velocity measurement, and have also failed to effectively resolve the contradiction between spatial coverage and temporal resolution.

[0009] Table 1 compares the method of the present invention with the prior art:

[0010] As can be seen from the table above, the existing technology has obvious shortcomings, and there is an urgent need for a modal analysis method for wind turbine towers that can fully leverage the advantages of FMCW lidar four-dimensional measurement. Summary of the Invention

[0011] To address the shortcomings of existing technologies, the present invention aims to provide a method and system for modal analysis of wind turbine towers based on a two-stage scanning approach using FMCW four-dimensional lidar. This invention fully utilizes the unique capability of FMCW lidar to directly measure radial velocity through Doppler frequency shift, and innovatively proposes a two-stage scanning strategy. The first stage employs a helical scan with oscillation compensation to rapidly acquire the overall geometric shape of the tower. The second stage uses an adaptive vertical reciprocating scan based on the geometric model to acquire high-density temporal velocity data, thereby resolving the technical problem of the contradiction between spatial coverage and temporal resolution in existing technologies.

[0012] Technical solution A modal analysis method for wind turbine towers based on two-stage scanning of FMCW four-dimensional lidar includes the following steps: S1: The first stage of the overall helical scan of the wind turbine tower is performed using an FMCW lidar. The FMCW lidar operates based on the principles of frequency-modulated continuous wave transmission and Doppler frequency shift detection, simultaneously outputting three-dimensional spatial coordinates (x, y, z) and radial velocity (v) for each measurement point, forming four-dimensional point cloud data. The first stage scan uses a helical path with oscillation compensation to quickly acquire the first four-dimensional point cloud data P1 of the overall geometric shape of the tower.

[0013] The helical scan path generation algorithm is as follows:

[0014]

[0015] in, and These represent the vertical and horizontal perspectives at time t, respectively. T1 is the maximum vertical elevation angle; T1 is the scan duration; N1 is the number of spiral turns; This refers to the range of horizontal scanning angles.

[0016] This invention proposes an oscillation compensation term. In conventional helical scanning, the scan line spacing at the top of the tower (high elevation angle region) increases with the elevation angle, resulting in insufficient top coverage. The oscillation compensation term causes the laser beam to periodically jitter in the vertical direction while rising along the helix, effectively filling the gaps between the helical lines, particularly improving coverage in the high elevation angle region.

[0017] The parameter optimization criterion for the oscillation compensation term is:

[0018]

[0019] in This is the amplitude coefficient (values ​​range from 0.3 to 0.8). This represents the frequency coefficient (values ​​3-6). The derivation of the above optimization criterion is based on the following analysis: the angular distance between two adjacent rotations of the helical scan is approximately... The oscillation amplitude A should be proportional to the spacing to effectively fill the gap; the oscillation frequency f should be such that the oscillation is completed during each helical scan. A complete oscillation is performed to ensure a uniform distribution of the compensation effect.

[0020] S2: Based on the first four-dimensional point cloud data P1, identify the geometric features of the tower and perform the second stage of vertical line reciprocating scanning.

[0021] First, spatial coordinate components are extracted from P1, and the RANSAC cylindrical fitting algorithm is used to identify the direction of the central axis of the tower and the range of the surface boundary.

[0022] Then, the second-stage scanning parameters are adaptively determined based on the results of the first-stage scan and environmental conditions: Determining the number of vertical scan lines N2: Calculate the point cloud coverage in the top region of the tower (height > 80% of tower height) during the first stage. Adjustments will be made according to the following rules:

[0023] in This is the coverage threshold (typically 90%). This is the number of baseline scan lines (typically 8).

[0024] Determining the number of reciprocating scans M for each vertical line:

[0025] Where v is the on-site wind speed. Reference wind speed (typical value 15m / s) This represents the maximum number of reciprocating cycles (typically 8 times). The physical meaning of this formula is: at low wind speeds, the tower vibration amplitude is small, requiring more scans to accumulate sufficient vibration cycle data; at high wind speeds, the vibration amplitude is large and the signal is obvious, and to reduce the impact of strong wind loads on scanning stability, the number of scans should be appropriately reduced.

[0026] Finally, N2 vertical scanning lines are uniformly planned on the surface of the tower, and M reciprocating scans are performed on each line to obtain second four-dimensional point cloud data P2 containing high-density temporal information.

[0027] S3: Four-dimensional point cloud data processing and temporal velocity matrix construction.

[0028] (1) Point cloud registration and fusion: P1 and P2 are registered based on spatial coordinates using the ICP (Iterative Closest Point) algorithm to obtain a complete four-dimensional point cloud P in a unified coordinate system.

[0029] (2) Tower segmentation and layering: The RANSAC algorithm is used to segment and extract the tower point cloud from P, and it is divided into K layers along the height direction.

[0030] (3) Construction of the temporal velocity matrix: For each time step For each level k, extract the 4D point cloud subset for that level at that time. The radial velocity values ​​measured and output by the FMCW lidar at each point are directly read, and the representative velocity value of this level is calculated using the weighted median method. :

[0031] Weight The calculation formula is:

[0032]

[0033]

[0034] in For point The angle of view upwards, This is the angle weight attenuation coefficient (typical value 20°). Reflection intensity, This represents the maximum reflection intensity at the current layer.

[0035] The design basis of angle weighting: FMCW lidar measures the radial velocity in the line of sight direction. When the line of sight is more perpendicular to the tower surface, the radial velocity can more accurately reflect the horizontal vibration of the tower. When the angle between the line of sight and the tower surface is too small, the measured radial velocity mainly reflects the vertical component, and the sensitivity to horizontal vibration is reduced.

[0036] The design basis for intensity weighting is that the higher the reflection intensity, the better the quality of the laser echo signal and the more reliable the measurement results.

[0037] The reason for using the median instead of the mean is that there are often outliers in lidar point clouds that are affected by atmospheric particle scattering and edge effects. The median statistic is robust to outliers and can effectively suppress the interference of noise on velocity extraction.

[0038] Finally, construct the time-series velocity matrix. Where K is the number of levels and N is the number of time steps.

[0039] S4: Modal parameter extraction.

[0040] (1) Natural frequency identification: Perform FFT (Fast Fourier Transform) on each row (each level) of the time-series velocity matrix V, calculate the power spectral density PSD, and identify the natural frequencies f_1, f_2, ..., f_n of each mode through peak detection.

[0041] (2) Damping ratio calculation: An adaptive weighted fusion of the half-power bandwidth method and the logarithmic decay method is adopted.

[0042] Half-power bandwidth damping ratio: ,in The resonant frequency, and This is the frequency corresponding to the half-power point.

[0043] Logarithmic decay damping ratio: ,in Let A be the logarithmic decay rate and A be the vibration amplitude envelope.

[0044] Adaptive determination of weighting coefficient α:

[0045] Where C_HPB is the confidence level of the half-power bandwidth method (determined by the signal-to-noise ratio of the peak spectral value), and C_LD is the confidence level of the logarithmic attenuation method (determined by the goodness of fit R² of the attenuation envelope).

[0046] (3) Modal shape extraction: Perform SVD (singular value decomposition) on the time-series velocity matrix V. The left singular vector corresponds to the relative amplitude distribution of each mode shape at different height levels.

[0047] S5 (Optional): Quality Assessment and Feedback. Calculate the Modal Assurance Criterion (MAC) matrix to assess the independence of each mode. If the maximum value of the off-diagonal elements of the MAC matrix exceeds the threshold (typically 0.2), or the coefficient of variation of the intrinsic frequencies exceeds the threshold (typically 5%), adjust the scan parameters and return to S1 to re-execute.

[0048] Beneficial effects Compared with the prior art, the present invention has the following significant advantages: 1. Fully leverage the advantages of FMCW four-dimensional measurement to achieve high-precision velocity acquisition. This invention directly utilizes the radial velocity measured by FMCW lidar based on the Doppler frequency shift principle, eliminating the need for indirect calculation through displacement differential. Compared with differential methods: (a) it eliminates the noise amplification effect introduced by differential calculation, improving velocity measurement accuracy from 0.1-0.2 m / s to 0.03-0.05 m / s; (b) velocity data and position data are acquired synchronously, and the time resolution is not limited by the differential interval; (c) it avoids the time delay of differential calculation, supporting real-time processing.

[0049] 2. An innovative two-stage scanning strategy resolves the contradiction between spatiotemporal resolution. The first stage, a helical scan with oscillation compensation, achieves over 90% spatial coverage within 15-30 seconds, establishing a precise geometric reference. The second stage, based on the geometric reference, performs targeted vertical reciprocating scans to acquire high temporal density vibration data. The total duration of the two stages is controlled to approximately one minute, improving efficiency by over 80% compared to traditional single-mode scanning.

[0050] 3. Oscillation compensation term optimizes the helical scan path, improving top coverage. Addressing the issue of sparse line spacing in conventional helical scans at high elevation angles, this invention introduces an oscillation compensation term into the helical path and provides parameter optimization criteria. Experiments show that after applying oscillation compensation, the coverage of the tower top (height > 80% of tower height) increases from 75% to 92%, significantly improving the accuracy of identifying higher-order mode shapes.

[0051] 4. Adaptive scanning parameter adjustment mechanism to adapt to different working conditions. The number of scan lines in the second stage is dynamically adjusted based on the coverage of the first stage, and the number of reciprocating strokes is adaptively adjusted based on wind speed to ensure optimal data quality under different environmental conditions.

[0052] 5. Robust velocity extraction algorithm based on weighted median. A weighting factor is designed considering both line-of-sight angle and reflection intensity, and median statistics are used to suppress outlier noise, effectively improving the signal-to-noise ratio of the velocity matrix.

[0053] 6. Improved recognition accuracy by fusing damping ratio calculation methods. Adaptively fusing the half-power bandwidth method and the logarithmic decay method, and allocating weights according to confidence level, the damping ratio recognition accuracy is improved by more than 30% compared to the single method, with the relative error controlled within 8%. Attached Figure Description

[0054] Figure 1 This is an overall flowchart of the method of the present invention; Figure 2 This is a schematic diagram of a two-stage scanning path; Figure 3 This is a three-dimensional schematic diagram of the overall spiral scanning path, showing the scanning trajectory with oscillation compensation. Figure 4 This is a schematic diagram of the reciprocating scanning path of a vertical line; Figure 5 A schematic diagram of point cloud registration and tower segmentation; Figure 6 A schematic diagram of the construction of the time-series velocity matrix is ​​provided, illustrating the method for calculating the weighted median. Figure 7 The flowchart for modal parameter extraction illustrates the frequency domain analysis and damping ratio fusion calculation process; Figure 8 The power spectral density diagram is shown in the embodiment; Figure 9 This is the modal shape diagram in the embodiment. Detailed Implementation

[0055] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0056] FMCW 4D LiDAR Working Principle The FMCW lidar used in this invention operates based on the principle of frequency-modulated continuous wave coherent detection, and its measurement process is as follows: (1) Signal emission: The laser emits a continuous laser signal whose frequency changes linearly with time, and the frequency modulation law is as follows: ,in B is the starting frequency, and B is the frequency modulation bandwidth. This is the frequency modulation period.

[0057] (2) Signal reception: After the laser shines on the target and is reflected, it takes a round trip time. (R is the target distance, c is the speed of light) and is then received. Due to propagation delay, there is a frequency difference between the received signal and the transmitted signal.

[0058] (3) Distance measurement: The transmitted signal and the received signal are mixed to generate a beat frequency signal, and the beat frequency... The distance is proportional to the target distance; the distance can be accurately calculated by analyzing the beat frequency using FFT.

[0059] (4) Velocity measurement: When the target is in radial motion, the reflected light will produce a Doppler frequency shift. (v is the radial velocity, λ is the laser wavelength). FMCW lidar can directly obtain the radial velocity of the target by analyzing the Doppler component in the beat frequency signal, without the need for differential calculation.

[0060] (5) Four-dimensional output: Integrating distance measurement, angle encoding and velocity measurement, four-dimensional data (x,y,z,v) are output synchronously for each measurement point.

[0061] Compared to differential velocimetry using traditional lidar, FMCW direct velocimetry has fundamental advantages: differential velocimetry has lower noise variance. When Δt decreases, the noise increases sharply; while the accuracy of FMCW direct velocity measurement depends only on the resolution of the Doppler frequency shift and is independent of the sampling interval, thus achieving both high accuracy and high time resolution.

[0062] Example 1: Modal Analysis of a 100-meter Onshore Wind Turbine Tower This embodiment performs modal analysis on a 100-meter-high steel-concrete hybrid tower to verify the effectiveness of the method of the present invention.

[0063] S1: Equipment Layout and First-Stage Scan An FMCW four-dimensional lidar system was installed 100 meters away from the tower. Equipment parameters: laser wavelength 1550nm, frequency modulation bandwidth 300MHz, frequency modulation period 100μs, sampling rate 8000 points / second, radial velocity measurement accuracy 0.04m / s, and ranging accuracy 3mm.

[0064] First-stage scanning parameters: Maximum vertical elevation angle Horizontal scanning range The number of spiral turns N1=4, and the scan duration T1=20 seconds. The oscillation compensation parameters are calculated according to the optimization criteria: k_a=0.5, k_f=4, resulting in A=45° / 4×0.5=5.6°≈6°, f=4×4 / 20=0.8Hz×4=3.2Hz≈3Hz.

[0065] In the first phase, approximately 160,000 four-dimensional point cloud data points were collected, with each point containing (x_i, y_i, z_i, v_i). Statistically, the overall tower coverage η = 91.3%, and the coverage of the top area (80-100m) η_top = 88.5%.

[0066] S2: Second-stage adaptive scanning Based on the first-stage point cloud identification, the tower's geometric features are: the central axis deviates from the vertical direction by 0.3°, the bottom diameter is 4.5m, and the top diameter is 2.8m.

[0067] The on-site wind speed measurement is v = 10 m / s. Adaptive parameter calculation: because , ; .

[0068] In the second stage, eight vertical scanning lines were evenly arranged on the surface of the tower. Each line was scanned four times, with a scanning time of about 40 seconds, and a total of about 320,000 four-dimensional point cloud data were collected.

[0069] S3: Data Processing and Speed ​​Matrix Construction ICP algorithm was used to register P1 and P2, with a registration RMSE of 0.7 cm. RANSAC algorithm was used to segment the tower point cloud, dividing it into K=10 layers along the height direction, each layer being 10 meters high.

[0070] Temporal velocity matrix construction: total sampling time 60 seconds, time resolution 0.02 seconds (50Hz), time steps N=3000. Angle weight attenuation coefficient. Calculate the weighted median velocity for each layer at each time step, and obtain... .

[0071] S4: Modal parameter extraction FFT analysis was performed on V, and the power spectral density plot was obtained. Figure 8 The three peaks are clearly displayed, and the natural frequencies of the first three modes are identified: f1=0.254Hz, f2=1.127Hz, and f3=3.456Hz.

[0072] Damping ratio calculation: First-order mode half-power bandwidth method Logarithmic decay method Confidence level C_HPB=0.85, C_LD=0.72, α=0.54, combined damping ratio Similarly, calculations yielded... , .

[0073] Modal shape extraction: The first three modal shapes are obtained by SVD decomposition ( Figure 9 The first mode is a first-order bending mode with a peak amplitude normalized to 1.00; the second mode is a second-order bending mode with a node at a height of approximately 60m; and the third mode is a third-order bending mode.

[0074] S5: Quality Assessment The MAC matrix calculation results show that all diagonal elements are greater than 0.98, and the maximum value of off-diagonal elements is 0.08, which satisfies the modal independence requirement.

[0075] Comparison with a synchronously installed reference accelerometer verified that the relative error of frequency was <0.5% (f1 error 0.001Hz), the relative error of damping ratio was <8%, and the mode shape MAC value was >0.95, thus verifying the accuracy of this method.

[0076] Example 2: Modal Analysis of a 120-meter Offshore Wind Turbine Tower This embodiment focuses on a 120-meter-high offshore wind turbine tower, highlighting the adaptability of the invention to harsh environments and the advantages of FMCW direct velocity measurement compared to differential velocity measurement.

[0077] Equipment parameters: laser wavelength 1550nm, sampling rate 10000 points / second, speed measurement accuracy 0.03m / s. Installation distance 80 meters.

[0078] First-stage scan parameters: , N1=5, T1=25 seconds, oscillation parameters A=4°, f=4Hz. Obtain coverage. , .

[0079] because Adaptively add N2=10 lines. On-site wind speed v=8m / s, M=8×exp(-8 / 15)=4.7≈5 scans. Second-stage scan duration: 55 seconds.

[0080] To compare the performance of FMCW direct speed measurement and traditional differential speed measurement, both methods were used to process the data simultaneously: Method A (this invention): The time-series velocity matrix is ​​constructed directly using the radial velocity v output by the FMCW lidar.

[0081] Method B (Differential Comparison): Using only spatial coordinates (x, y, z), through difference... Calculation speed .

[0082] The comparison results are as follows: (1) Velocity signal-to-noise ratio: The velocity time series signal-to-noise ratio of method A is 32dB, while that of method B is 18dB, an improvement of 14dB.

[0083] (2) Frequency identification: Method A identifies f1=0.231Hz, Method B identifies f1=0.229Hz, and the reference value is 0.230Hz; Method A has an error of 0.4%, Method B has an error of 0.4%, and the two are comparable.

[0084] (3) Damping ratio identification: Calculated by method A Method B has a relative error of 1.15%, while the reference value is 0.85%; Method A has a relative error of 2.4%, and Method B has a relative error of 35%. The damping ratio is sensitive to noise, and the advantages of FMCW direct speed measurement are fully demonstrated here.

[0085] (4) High-order mode recognition: Method A successfully recognized the first 4 modes, while Method B could only reliably recognize the first 2 modes. The 3rd and 4th modes were submerged in differential noise.

[0086] This embodiment verifies the significant advantages of FMCW direct velocity measurement over traditional differential velocity measurement in damping ratio identification and higher-order mode identification.

[0087] Example 3: Long-term monitoring under low wind speed conditions This embodiment focuses on modal monitoring under low wind speed conditions (v<5m / s) to illustrate the role of the adaptive parameter adjustment mechanism.

[0088] The target tower is 90 meters high and the erection distance is 120 meters. The on-site wind speed is v = 3 m / s.

[0089] The first phase uses standard parameter scanning, coverage... , .

[0090] Because the wind speed is very low, the adaptive calculation M = 8 × exp(-3 / 15) = 8 × 0.82 = 6.5 ≈ 7 times, N2 = 8 lines (because...). No need to add).

[0091] The increased number of reciprocating cycles extended the data acquisition time to 80 seconds, but ensured sufficient accumulation of vibration period data under low amplitude conditions. Analysis results showed that the frequency variation coefficients for the first three modes were <2%, and the damping ratio variation coefficients were <10%, indicating that the data quality met the requirements.

[0092] If fixed parameters (M=4) are used, the frequency variation coefficient reaches 5% and the damping ratio variation coefficient reaches 25% under the same conditions, which cannot meet the engineering accuracy requirements. This verifies the necessity of an adaptive parameter adjustment mechanism.

[0093] The above description is merely a preferred embodiment of the present invention, and the present invention is not limited to the above embodiments. It is understood that other improvements and variations that are directly derived or conceived by those skilled in the art without departing from the spirit and concept of the present invention should be considered to be included within the protection scope of the present invention.

Claims

1. A modal analysis method for wind turbine towers based on two-stage scanning of FMCW four-dimensional lidar, characterized in that, Includes the following steps: S1: The FMCW lidar is used to perform the first stage of overall spiral scanning of the wind turbine tower. The FMCW lidar is based on the principle of frequency-modulated continuous wave transmission and Doppler frequency shift detection. It synchronously outputs three-dimensional spatial coordinates and radial velocity for each measurement point to obtain the first four-dimensional point cloud data of the overall geometric shape of the tower. S2: Based on the first four-dimensional point cloud data, identify the geometric features of the tower, perform the second stage of vertical line reciprocating scanning, and obtain the second four-dimensional point cloud data containing high-density temporal information; S3: Register and fuse the first and second four-dimensional point cloud data, perform layered processing on the tower, directly extract the radial velocity values ​​measured by FMCW lidar in each level of four-dimensional point cloud, and construct a time-series velocity matrix. S4: Perform frequency domain analysis on the time-series velocity matrix to extract the natural frequency, damping ratio, and mode shape of the tower.

2. The method according to claim 1, characterized in that, The radial velocity measurement of the FMCW lidar is based on the following principle: a continuous laser signal with a frequency linearly modulated over time is emitted, and an echo signal reflected from the target is received. The frequency change of the echo signal relative to the emitted signal contains two components: a range-dependent beat frequency component caused by the optical path difference and a Doppler frequency shift component caused by the target motion. The two components are separated through signal processing, and the target distance and radial velocity are obtained synchronously.

3. The method according to claim 1, characterized in that, The path generation algorithm for the first stage of the overall spiral scan is as follows: in, Let be the vertical perspective at time t. The horizontal perspective at time t The maximum vertical elevation angle is T1, the scan duration is N1, and the number of spiral turns is N1. This is an oscillation compensation term used to compensate for the sparse line spacing problem in the high elevation angle region of helical scanning, where A is the oscillation amplitude and f is the oscillation frequency; the parameters of the oscillation compensation term are determined according to the following optimization criteria: Wherein, k_a is the amplitude coefficient, with a value range of 0.3-0.8; k_f is the frequency coefficient, with a value range of 3-6; the optimization criterion improves the uniformity of the projection point distribution of the oscillation-compensated scanning trajectory on the tower surface.

4. The method according to claim 1, characterized in that, The parameters for the second-stage vertical reciprocating scan are adaptively adjusted based on the scan results of the first stage and environmental conditions: The method for determining the number of vertical scan lines N2 is as follows: calculate the point cloud coverage rate of the first-stage scan in the top region of the tower. ,like ,but ;like ,but ;in The coverage threshold, The number of baseline scan lines; The method for determining the number of reciprocating scans M for each vertical line is as follows: Where v is the on-site wind speed, For reference wind speed, The maximum number of reciprocating cycles is determined by the method, which allows for the acquisition of more vibration cycle data at low wind speeds and reduces scanning time at high wind speeds to mitigate the impact of wind load.

5. The method according to claim 1, characterized in that, The method for constructing the time-series velocity matrix in step S3 includes: The tower is divided into K levels along its height; for each time step Extract the four-dimensional point cloud subset at the k-th level. ; Calculate the representative velocity value for this level. The weighted median method is used: in For point Radial velocity directly measured by FMCW lidar This represents the weight of that point; The weight Determined by the product of the angle weight component and the intensity weight component: in For point The angle of view upwards, This is the angle weight attenuation coefficient. For point The intensity of reflection, The maximum reflection intensity of the current level; the angle weight increases the weight of measurement points whose line of sight is more perpendicular to the tower surface, and the intensity weight increases the weight of measurement points with better signal quality.

6. The method according to claim 1, characterized in that, In step S4, the damping ratio is calculated using an adaptive weighted fusion of the half-power bandwidth method and the logarithmic decay method. Where ζ_HPB is the damping ratio calculated using the half-power bandwidth method. The damping ratio is calculated using the logarithmic decay method; The weighting coefficient α is adaptively determined based on the confidence levels of the two methods: Where C_HPB is the confidence level of the half-power bandwidth method, which is determined by the signal-to-noise ratio of the peak spectral value; The confidence level of the logarithmic decay method is determined by the goodness of fit of the decay envelope.

7. The method according to claim 1, characterized in that, The operating parameters of the FMCW lidar are: laser wavelength 1500-1600nm, frequency modulation bandwidth 200-500MHz, sampling rate 5000-15000 points / second, and radial velocity measurement accuracy better than 0.1m / s; the parameter range for the first stage of overall helical scanning is: maximum vertical elevation angle The horizontal scanning angle range is 40-50°. The range is ±20° to ±40°, the number of spiral turns N1 is 3-6 turns, and the scan duration T1 is 15-30 seconds.

8. The method according to claim 1, characterized in that, It also includes step S5: assess the quality of the modal analysis results, calculate the modal guarantee criterion MAC matrix, and if the maximum value of the off-diagonal elements of the MAC matrix exceeds the preset threshold, or the coefficient of variation of the intrinsic frequency exceeds the preset threshold, then adjust the scanning parameters and return to step S1 to re-execute.

9. A four-dimensional point cloud data processing method for modal analysis of wind turbine towers, characterized in that, include: Receive four-dimensional point cloud data collected by FMCW lidar, wherein each point in the four-dimensional point cloud data contains three-dimensional spatial coordinates (x, y, z) and radial velocity v directly measured by Doppler frequency shift; Based on spatial coordinates, four-dimensional point cloud data of multiple time phases are registered to establish a unified spatial reference system; The tower region is segmented and extracted from the registered four-dimensional point cloud and layered along the tower height direction; For each level and each time step of the point cloud subset, the radial velocity values ​​are directly extracted, and the representative velocity values ​​are calculated using a weighted median method based on the line-of-sight angle and reflection intensity. Representative velocity values ​​from each level are organized in time series to construct a time-series velocity matrix for subsequent frequency domain modal analysis.

10. A wind turbine tower modal analysis system based on FMCW four-dimensional lidar dual-stage scanning, characterized in that, include: The FMCW lidar unit is equipped with a frequency-modulated continuous wave laser transmitter and a coherent receiver. It is used to transmit frequency-modulated continuous laser and receive the target reflected echo. It obtains the target distance through beat frequency analysis and the target radial velocity through Doppler frequency shift analysis. It outputs four-dimensional data for each measurement point. The scanning control unit includes a two-dimensional precision turntable and a motion controller, which is used to control the FMCW lidar unit to scan according to a preset two-stage scanning trajectory. The two-stage scanning trajectory includes a first-stage spiral scanning trajectory with oscillation compensation and a second-stage adaptive vertical reciprocating scanning trajectory. The data processing unit is used to perform registration and fusion of the acquired four-dimensional point cloud data, tower segmentation and layering, and construction of a time-series velocity matrix based on direct velocity measurement by FMCW lidar. The modal analysis unit is used to perform frequency domain transformation, natural frequency identification, damping ratio calculation, and mode shape extraction on the time-series velocity matrix.

Images