A scour monitoring method for offshore wind power jacket foundation based on dynamic characteristics

Abstract

The application discloses a kind of based on power feature offshore wind power jacket foundation scour monitoring method, comprising: obtaining geological exploration data, fan design parameters, by finite element construction 1 order generalized flexibility matrix;Theoretical 1 order generalized flexibility matrix of benchmark condition is calculated;Theoretical 1 order generalized flexibility matrix corresponding to each condition is calculated;The generalized flexibility change rate of each measuring point is calculated;The generalized flexibility change rate original data of each measuring point is converted into three-dimensional feature vector of full field mean value, single pile mean value, skewness;Three-dimensional feature vector is used as input, and the scour depth prediction model based on bayesian regularization algorithm is established;Measured acceleration signal is collected, and 1 order generalized flexibility matrix under measured condition is obtained by reconstruction, and measured three-dimensional feature vector is converted from measured matrix;Measured three-dimensional feature vector is input into the scour depth prediction model that has been trained, and then the predicted scour depth value of each pile is output.The application realizes the accurate prediction of each leg scour depth.

Description

Technical Field

[0001] This invention relates to a method for monitoring the scour of offshore wind turbine jacket foundations based on dynamic characteristics. Background Technology

[0002] In the field of marine resource development and space utilization, marine pile foundations serve as the core foundation supporting superstructures, and their stability is crucial for ensuring the safe operation of energy platforms, subsea data centers, and monitoring stations. However, the presence of pile foundations significantly disturbs the local flow field, inducing complex eddy evolution and altering water flow shear stress, disrupting the original sediment transport balance, and leading to localized scour. Localized scour not only reduces the effective burial depth of the pile foundation but also increases the bending moment generated by horizontal loads, resulting in weakened pile-soil interaction and significant degradation of horizontal bearing capacity, seriously threatening the overall stability and fatigue life of the structure.

[0003] Efficient scour monitoring is a prerequisite for implementing precise protection and control. Currently, while traditional embedded sensors, sonar, and depth sounding rod technologies are relatively mature in inland river bridge pier monitoring, their installation and maintenance costs are high, sensors are easily damaged, and measurement accuracy is difficult to guarantee in the high-salt, high-pressure, and strongly dynamic environments of the deep sea. While the existing multibeam scanning technology in the offshore wind power sector can provide high-precision three-dimensional topography, its high operating costs limit its application to periodic inspections, making real-time monitoring of the dynamic scour process impossible. Furthermore, while monitoring methods based on structural dynamic characteristics have holistic advantages, research on jacket foundations is insufficient, especially under complex conditions with multiple coupled factors, making accurate prediction of the scour depth of each leg difficult and exhibiting significant monitoring limitations.

[0004] Therefore, to address the above issues, a method for monitoring the scour of offshore wind turbine jacket foundations based on dynamic characteristics is provided. Summary of the Invention

[0005] To address the aforementioned problems in the existing technology, this invention provides a method for monitoring the scour of offshore wind turbine jacket foundations based on dynamic characteristics, which enables accurate prediction of the scour depth of each leg.

[0006] The technical solution to achieve the above objectives is: A method for monitoring scour of offshore wind turbine jacket foundations based on dynamic characteristics, comprising: Step S1: Obtain geological survey data and wind turbine design parameters. Use finite element software to establish a wind turbine-jacket-pile-soil interaction model. Obtain the natural frequency and mode shape matrix of the structure through finite element calculation. Take the first three modes to construct the first-order generalized flexibility matrix. Step S2: Define the no-scour condition as the baseline condition and calculate the theoretical first-order generalized compliance matrix of the baseline condition. ; Step S3: Set up no fewer than 25 asynchronous scour simulation conditions, extract the generalized displacement response at the sensor placement nodes under each condition, and calculate the theoretical first-order generalized compliance matrix corresponding to each condition. ; Step S4: Calculate the theoretical first-order generalized compliance matrix for each scouring condition. Theoretical first-order generalized compliance matrix of the reference working condition Compare and calculate the generalized compliance rate of change at each measuring point. ; Step S5: For each working condition, calculate the generalized compliance rate of change at each measuring point. The raw data is transformed into a three-dimensional feature vector of overall average, single pile average, and skewness; Step S6: Take the three-dimensional feature vector as input and the corresponding preset scour depth as output label to establish a scour depth prediction model based on Bayesian regularization algorithm. Step S7: Through iterative training, the scour depth prediction model learns the nonlinear mapping relationship from the spatial flexibility heterogeneity characteristics to the independent scour depth of each pile. Step S8: Acquire the measured acceleration signal and combine it with the reduced-order mass matrix of the reference finite element model. By modal superposition, the first-order generalized compliance matrix under the measured working conditions is reconstructed. And according to the logic of steps S3 and S4, the measured matrix is ​​transformed into a measured three-dimensional feature vector; Step S9: Input the measured three-dimensional feature vector into the trained scour depth prediction model, and then output the average scour depth value of the entire offshore wind turbine jacket, or the predicted scour depth value of each pile.

[0007] Preferably, in step S1, the formula for calculating the first-order generalized compliance matrix is ​​as follows: ; In the formula, and The first The first natural frequency and the corresponding mode shape vector, This is the transpose symbol.

[0008] Preferably, in step S3, the sensor arrangement principle is that the four piles of the offshore wind turbine jacket are arranged symmetrically, with one sensor placed at the mud surface and one at the top of each pile, for a total of 8 sensors arranged in the entire structure.

[0009] Preferably, in step S5, the generalized compliance change rate of each measuring point is... The raw data is transformed into a three-dimensional feature vector comprising the overall average, the average value of a single pile, and the skewness, including: The generalized compliance rate of change is obtained by subtracting the diagonal elements of the first-order generalized compliance matrix before and after scouring. : ; Suppose that under a certain working condition, the total number of [unclear] is [unclear]. The generalized compliance change rate at each of the eight sensor locations is... The set is: ; The four piles of the offshore wind turbine jacket are designated as pile a, pile b, pile c, and pile d. For the piles numbered... The piles, among which, The corresponding set of node response values ​​is: ; when hour, Take respectively ; Average of the whole game The average scour level, reflecting the overall structure, serves as a background benchmark and is expressed as follows: ; pile average The expression reflecting the local structural damage strength of a single pile is as follows: ; The average second central moment of the four piles and third-order center The distances are respectively: ; ; Skewness It is the ratio of the third central distance to the cube of the standard deviation, expressed as follows: ; when If the scouring is consistent across the four piles, it indicates that the scouring is symmetrically distributed. when If the value is zero, it indicates that the scouring degree of the four piles is inconsistent. The greater the difference between this value and 0, the more unbalanced the scouring development of the four piles.

[0010] Preferably, in step S5, The average value of the whole game Accurate prediction of the average scour depth of four piles using quadratic regression: ; In the formula, , and All are regression coefficients. Predicted value of average scour depth.

[0011] Preferably, in step S8, the specific reconstruction process of the measured first-order generalized compliance matrix is ​​as follows: Using synchronous acquisition The road acceleration response signal is used to identify the first three natural frequencies of the structure through an environmental modal recognition algorithm. and relative mode vector ,in, Introducing a reduced-order mass matrix The normalized measured first three vibration modes were obtained. : ; Based on the first three normalized measured mode shapes and frequencies, the measured first-order generalized compliance matrix is ​​reconstructed. : .

[0012] Compared with the prior art, the beneficial effects of the present invention are: 1) A non-contact accelerometer is used, and modal mode inversion is used instead of direct underwater measurement, which avoids the sealing failure problem of traditional buried sensors in water depths >50m; 2) The 8-node layout scheme (mud surface + pile top) reconstructs the full-field displacement field through modal extension theory, which reduces the number of sensors and lowers the cost compared to full-pile strain monitoring; 3) A reference matrix is ​​established through a finite element model (wind turbine-jacket-pile-soil interaction) to eliminate the influence of seabed geological time-varying characteristics. The frequency band characteristics of wave and wind turbine operation are automatically identified through a Bayesian regularized neural network. The structural modal shift caused by scouring is separated in the frequency domain, which improves the signal-to-noise ratio. The superposition of the first three modes significantly improves the sensitivity to micro-scouring and improves the scouring recognition rate. The introduction of a reduced-order mass matrix solves the uncertainty of the modal scaling factor under environmental excitation and reduces the mode shape normalization error. Attached Figure Description

[0013] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of a method for monitoring the scour of offshore wind turbine jacket foundations based on dynamic characteristics, according to the present invention. Figure 2 This is another flowchart of the present invention, which is a method for monitoring the scour of offshore wind turbine jacket foundations based on dynamic characteristics; Figure 3 This is a schematic diagram of the offshore wind power jacket foundation of the present invention. Detailed Implementation

[0014] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0015] like Figure 1 , 2 As shown, a method for monitoring scour of offshore wind turbine jacket foundations based on dynamic characteristics includes: Step S1: Obtain geological survey data and wind turbine design parameters. Use finite element software to establish a wind turbine-jacket-pile-soil interaction model. Obtain the natural frequency and mode shape matrix of the structure through finite element calculation. Take the first three modes to construct the first-order generalized flexibility matrix.

[0016] In this embodiment, the formula for calculating the first-order generalized compliance matrix is ​​as follows: ; In the formula, and The first The first natural frequency and the corresponding mode shape vector, This is the transpose symbol.

[0017] Step S2: Define the no-scour condition as the baseline condition and calculate the theoretical first-order generalized compliance matrix of the baseline condition. .

[0018] Step S3: Set up no fewer than 25 asynchronous scour simulation conditions, extract the generalized displacement response at the sensor placement nodes under each condition, and calculate the theoretical first-order generalized compliance matrix corresponding to each condition. .

[0019] In this embodiment, the sensor arrangement principle is that the four piles of the offshore wind turbine jacket are symmetrically arranged, with one sensor placed at the mud surface and one at the top of each pile, for a total of eight sensors arranged in the entire structure. Figure 3 As shown, the offshore wind turbine jacket consists of jacket structure 2, tower 3, wind turbine blades 4, pile a 5, pile b 6, pile c 7, and pile d 8. A sensor 1 is arranged at the mud surface and the top of piles a 5, b 6, c 7, and d 8. The lower ends of piles a 5, b 6, c 7, and d 8 are fixed in the seabed 9, and the upper ends are fixedly installed with jacket structure 2. The tower 3 is fixed to the upper end of jacket structure 2, and the wind turbine blades 4 are fixed on the tower 3.

[0020] Step S4: Calculate the theoretical first-order generalized compliance matrix for each scouring condition. Theoretical first-order generalized compliance matrix of the reference working condition Compare and calculate the generalized compliance rate of change at each measuring point. .

[0021] Step S5: For each working condition, calculate the generalized compliance rate of change at each measuring point. The raw data is transformed into three-dimensional feature vectors of overall average, single pile average, and skewness.

[0022] In the embodiment, the generalized compliance change rate at each measuring point is... The raw data is transformed into a three-dimensional feature vector comprising the overall average, the average value of a single pile, and the skewness, including: The generalized compliance rate of change is obtained by subtracting the diagonal elements of the first-order generalized compliance matrix before and after scouring. : ; Suppose that under a certain working condition, the total number of [unclear] is [unclear]. The generalized compliance change rate at each of the eight sensor locations is... The set is: ; The four piles of the offshore wind turbine jacket are designated as pile a, pile b, pile c, and pile d. For the piles numbered... The piles, among which, The corresponding set of node response values ​​is: ; when hour, Take respectively ; Average of the whole game The average scour level, reflecting the overall structure, serves as a background benchmark and is expressed as follows: ; pile average The expression reflecting the local structural damage strength of a single pile is as follows: ; The average second central moment of the four piles and third-order center The distances are respectively: ; ; Skewness It is the ratio of the third central distance to the cube of the standard deviation, expressed as follows: ; when If the scouring is consistent across the four piles, it indicates that the scouring is symmetrically distributed. when If the value is zero, it indicates that the scouring degree of the four piles is inconsistent. The greater the difference between this value and 0, the more unbalanced the scouring development of the four piles.

[0023] In the embodiments, The average value of the whole game Accurate prediction of the average scour depth of four piles using quadratic regression: ; In the formula, , and All are regression coefficients. Predicted value of average scour depth.

[0024] Step S6: Using the three-dimensional feature vector as input and the corresponding preset scour depth as output label, a scour depth prediction model based on the Bayesian regularization algorithm is established.

[0025] Step S7: Through iterative training, the scour depth prediction model learns the nonlinear mapping relationship from the spatial flexibility heterogeneity characteristics to the independent scour depth of each pile.

[0026] Step S8: Acquire the measured acceleration signal and combine it with the reduced-order mass matrix of the reference finite element model. By modal superposition, the first-order generalized compliance matrix under the measured working conditions is reconstructed. Following the logic of steps S3 and S4, the measured matrix is ​​transformed into a measured three-dimensional feature vector.

[0027] In this embodiment, the specific reconstruction process of the measured first-order generalized compliance matrix is ​​as follows: Using synchronous acquisition The road acceleration response signal is used to identify the first three natural frequencies of the structure through an environmental modal recognition algorithm. and relative mode vector ,in, Introducing a reduced-order mass matrix The normalized measured first three vibration modes were obtained. : ; Based on the first three normalized measured mode shapes and frequencies, the measured first-order generalized compliance matrix is ​​reconstructed. : .

[0028] Step S9: Input the measured three-dimensional feature vector into the trained scour depth prediction model, and then output the average scour depth value of the entire offshore wind turbine jacket, or the predicted scour depth value of each pile.

[0029] Finally, it should be noted that the above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for monitoring scour of offshore wind turbine jacket foundations based on dynamic characteristics, characterized in that, include: Step S1: Obtain geological survey data and wind turbine design parameters. Use finite element software to establish a wind turbine-jacket-pile-soil interaction model. Obtain the natural frequency and mode shape matrix of the structure through finite element calculation. Take the first three modes to construct the first-order generalized flexibility matrix. Step S2: Define the no-scour condition as the baseline condition and calculate the theoretical first-order generalized compliance matrix of the baseline condition. ; Step S3: Set up no fewer than 25 asynchronous scour simulation conditions, extract the generalized displacement response at the sensor placement nodes under each condition, and calculate the theoretical first-order generalized compliance matrix corresponding to each condition. ; Step S4: Calculate the theoretical first-order generalized compliance matrix for each scouring condition. Theoretical first-order generalized compliance matrix of the reference working condition Compare and calculate the generalized compliance rate of change at each measuring point. ; Step S5: For each working condition, calculate the generalized compliance rate of change at each measuring point. The raw data is transformed into a three-dimensional feature vector of overall average, single pile average, and skewness; Step S6: Take the three-dimensional feature vector as input and the corresponding preset scour depth as output label to establish a scour depth prediction model based on Bayesian regularization algorithm. Step S7: Through iterative training, the scour depth prediction model learns the nonlinear mapping relationship from the spatial flexibility heterogeneity characteristics to the independent scour depth of each pile. Step S8: Acquire the measured acceleration signal and combine it with the reduced-order mass matrix of the reference finite element model. By modal superposition, the first-order generalized compliance matrix under the measured working conditions is reconstructed. And according to the logic of steps S3 and S4, the measured matrix is ​​transformed into a measured three-dimensional feature vector; Step S9: Input the measured three-dimensional feature vector into the trained scour depth prediction model, and then output the average scour depth value of the entire offshore wind turbine jacket, or the predicted scour depth value of each pile.

2. The method for monitoring scour of offshore wind turbine jacket foundations based on dynamic characteristics according to claim 1, characterized in that, In step S1, the formula for calculating the first-order generalized compliance matrix is ​​as follows: ； In the formula, and The first The first natural frequency and the corresponding mode shape vector, This is the transpose symbol.

3. The method for monitoring scour of offshore wind turbine jacket foundations based on dynamic characteristics according to claim 1, characterized in that, In step S3, the sensor arrangement principle is that the four piles of the offshore wind turbine jacket are arranged symmetrically, with one sensor placed on the mud surface and one on the top of each pile, for a total of eight sensors arranged in the entire structure.

4. The method for monitoring scour of offshore wind turbine jacket foundations based on dynamic characteristics according to claim 3, characterized in that, In step S5, the generalized compliance change rate of each measuring point is... The raw data is transformed into a three-dimensional feature vector comprising the overall average, the average value of a single pile, and the skewness, including: The generalized compliance rate of change is obtained by subtracting the diagonal elements of the first-order generalized compliance matrix before and after scouring. : ； Suppose that under a certain working condition, the total number of [unclear] is [unclear]. The generalized compliance change rate at each of the eight sensor locations is... The set is: ； The four piles of the offshore wind turbine jacket are designated as pile a, pile b, pile c, and pile d. For the piles numbered... The piles, among which, The corresponding set of node response values ​​is: ； when hour, Take respectively ; Average of the whole game The average scour level, reflecting the overall structure, serves as a background benchmark and is expressed as follows: ； pile average The expression reflecting the local structural damage strength of a single pile is as follows: ； The average second central moment of the four piles and third-order center The distances are respectively: ； ； Skewness It is the ratio of the third central distance to the cube of the standard deviation, expressed as follows: ； when If the scouring is consistent across the four piles, it indicates that the scouring is symmetrically distributed. when If the value is zero, it indicates that the scouring degree of the four piles is inconsistent. The greater the difference between this value and 0, the more unbalanced the scouring development of the four piles.

5. The method for monitoring scour of offshore wind turbine jacket foundations based on dynamic characteristics according to claim 4, characterized in that, In step S5 The average value of the whole game Accurate prediction of the average scour depth of four piles using quadratic regression: ； In the formula, , and All are regression coefficients. Predicted value of average scour depth.

6. The method for monitoring scour of offshore wind turbine jacket foundations based on dynamic characteristics according to claim 1, characterized in that, In step S8, the specific reconstruction process of the measured first-order generalized compliance matrix is ​​as follows: Using synchronous acquisition The road acceleration response signal is used to identify the first three natural frequencies of the structure through an environmental modal recognition algorithm. and relative mode vector ,in, Introducing a reduced-order mass matrix The normalized measured first three vibration modes were obtained. : ； Based on the first three normalized measured mode shapes and frequencies, the measured first-order generalized compliance matrix is ​​reconstructed. : 。

Images