A smart monitoring method and system for the bearing capacity of offshore wind turbine pile foundations

Abstract

This invention discloses an intelligent monitoring method and system for the bearing capacity of offshore wind turbine pile foundations, relating to the field of offshore wind turbine pile foundation monitoring technology. It comprises two main components: pile foundation horizontal displacement calculation and early warning, and wind turbine structural frequency calculation and early warning. Starting from the interaction relationship between the wind turbine pile and the soil, it considers the scour-induced performance degradation of the pile foundation and establishes a model algorithm that considers the impact of scour on the bearing capacity of the pile foundation and the calculation of the wind turbine structural frequency. This allows for dynamic prediction and trend analysis of the offshore wind turbine structural performance, forming a closed-loop management system from data acquisition and real-time processing to decision output. Therefore, by employing the aforementioned intelligent monitoring method and system for the bearing capacity of offshore wind turbine pile foundations, through a logical closed loop of multi-dimensional monitoring, high-precision analysis, and intelligent early warning, the technical deficiencies of offshore wind turbine pile foundations—difficult to monitor, assess, and warn against—are systematically addressed.

Description

Technical Field

[0001] This invention relates to the field of offshore wind power pile foundation monitoring technology, and in particular to an intelligent monitoring method and system for the bearing capacity of offshore wind power pile foundations. Background Technology

[0002] my country possesses abundant offshore wind energy resources, and clean, efficient offshore wind power has become the core of low-carbon energy solutions for coastal areas. This industry not only possesses environmentally friendly attributes but also occupies a strategic position in the development of the marine economy and the enhancement of competitiveness in the new energy industry. Offshore wind turbines are widely used in complex marine environments due to their advantages such as simple foundation structure, high load-bearing capacity, and low settlement.

[0003] However, the large-diameter monopile foundations of offshore wind turbines must withstand horizontal cyclic loads from wind, waves, and ocean currents during operation. The resulting horseshoe-shaped eddies and wake effects in the soil around the piles lead to seabed erosion and the formation of settlement zones. This process directly causes a decrease in the horizontal bearing capacity of the pile foundation, a shift in its natural frequency, and even structural failure. Specifically, cyclic loads weaken the soil's resistance to erosion through a dual mechanism of cumulative soil damage and intensified vibration. Furthermore, the increased erosion depth reduces the foundation stiffness, causing the pile foundation's natural frequency to decrease and approach the resonance risk range. Therefore, conducting research on the impact of erosion on the natural frequency of offshore wind turbines and constructing a precise monitoring and early warning system are key paths to solving the challenges of safe operation and maintenance of deep-sea offshore wind turbines.

[0004] Currently, existing monitoring and early warning systems suffer from the following problems: Firstly, existing monitoring sources are relatively singular, focusing only on the scour depth of the soil and neglecting the bearing capacity of the pile foundation and the structural performance of the wind turbine after scour. Secondly, existing monitoring technologies suffer from insufficient monitoring accuracy and excessively long monitoring times. Furthermore, the processing of monitoring data is crucial for early warning of pile foundation performance risks, but a full-lifecycle, multi-scale, real-time digital twin system has not yet been established in the offshore wind power sector. Therefore, it is necessary to develop a software platform suitable for monitoring and early warning of scour and seepage in offshore wind turbine pile foundations to address the shortcomings of existing monitoring methods, the insufficient accuracy of digital twin systems, and the lack of attention to pile foundation bearing capacity and structural performance monitoring. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent monitoring method and system for the bearing capacity of offshore wind power pile foundations. Through the technical architecture of "multi-sensor collaborative monitoring - dual-engine algorithm analysis - digital twin early warning", it systematically breaks through the bottlenecks of traditional monitoring in terms of accuracy, timeliness and multi-parameter decoupling, and forms a full-chain solution from intelligent perception to digital mapping and then to intelligent decision-making.

[0006] To achieve the above objectives, the present invention provides an intelligent monitoring method and system for the bearing capacity of offshore wind turbine pile foundations, comprising:

[0007] S1. Real-time displacement and vibration data of offshore wind power pile foundations are collected using displacement sensors and acceleration sensors. The collected data is then processed to extract the measured natural frequency of the structure and the horizontal displacement of the pile foundation.

[0008] S2. Based on a dual-engine architecture driven by physical mechanisms and data, a prediction model for the horizontal bearing capacity of large-diameter steel piles for wind turbines is constructed. Using multi-source data fusion technology, the bearing capacity of the pile foundation and the structural status of the wind turbine are evaluated or warned through analysis of the horizontal displacement field of the pile foundation and identification of the structural dynamic characteristics. Among them, the warning adopts a three-level threshold warning mechanism.

[0009] Furthermore, in S2, the prediction model for the horizontal bearing capacity of large-diameter steel piles for wind turbines includes: a soil softening model based on initial static shear stress, a modified py curve, and simulation of foundation stiffness degradation under cyclic loading; dynamic calibration of relevant parameters involved in the model by inverting scour depth and damping ratio through Bayesian neural network, and optimization of the pile foundation natural frequency calculated by the model.

[0010] Furthermore, S2 includes using In-SAR technology to generate a phase difference map by interferometric processing of the received signal, and then converting the radar line-of-sight displacement into a horizontal displacement to invert the horizontal displacement of the pile foundation.

[0011] Furthermore, the inversion of pile foundation horizontal displacement includes obtaining the regional displacement trend of multiple pile foundations through regional macro-monitoring, and conducting single-point fine analysis by utilizing the difference in scattering characteristics between the pile foundation and the surrounding soil; it also includes constructing a time-series trend model based on several years of image sequences to identify abnormal displacement growth.

[0012] Furthermore, S2 includes fusing the measured horizontal displacement of the pile foundation with the horizontal displacement of the pile foundation retrieved using In-SAR technology using a Kalman filter algorithm for dual displacement verification; when the displacement deviation is <5%, it is included in the evaluation; otherwise, a three-level threshold early warning mechanism is triggered:

[0013] When the displacement deviation is between 5% and 10% or the natural frequency deviates from the design value by 8%, a first-level early warning is triggered. The regional displacement trend and the measured horizontal displacement of the pile foundation are integrated to generate a sensor health diagnosis report and simultaneously generate an early warning matrix for inelastic deformation of the pile foundation.

[0014] When the displacement exceeds the limit by 10%-20% or the natural frequency drops by 10%, a level two early warning is triggered, the redundancy verification of the backup sensor is initiated, and the UAV lidar is linked to scan the mud surface of the pile foundation to generate a three-dimensional model containing the displacement vector field and frequency characteristic curve. Combined with the digital model simulation, the stiffness degradation trend is predicted.

[0015] When the displacement exceeds the limit by 20% or the natural frequency drops by 15%, a level 3 warning is triggered. Based on high-frequency data from the accelerometer and lidar data from the UAV, the emergency reserve is activated, and a repair work order package containing an underwater emergency repair plan and a spare parts list is generated.

[0016] This invention also provides an intelligent monitoring system for the bearing capacity of offshore wind power pile foundations, comprising a collaboratively designed perception layer, edge layer, platform layer, and application layer, forming a closed-loop management system for the entire process from data acquisition to decision output;

[0017] The perception layer is equipped with displacement sensors and acceleration sensors to monitor pile foundations at different depths in real time and store the monitoring data in a storage bin. The data is then collected and relayed through a wireless transmission bin. Furthermore, it integrates In-SAR satellite data and UAV lidar data to construct a three-level monitoring network from point to surface. Each level uses different monitoring methods to achieve pile foundation data monitoring from point to surface and from micro to macro.

[0018] The edge layer uses an embedded server to preprocess the original signal in real time, extracts the natural frequency using Fourier transform, and performs dual-source verification of the displacement data by combining wavelet denoising and Kalman filtering algorithms.

[0019] At the platform layer, based on the prediction model of the horizontal bearing capacity of large-diameter steel piles of wind turbines, the frequency spectrum characteristic curves drawn according to the data limit values ​​in the input standard specifications and the natural frequency of the wind turbine structure are dynamically evaluated to assess the safety status of the pile foundation and predict the ultimate bearing capacity.

[0020] At the application layer, an intelligent early warning platform centered on a digital twin dashboard is set up, which integrates finite element models with real-time data, dynamically renders stress concentration areas, and is equipped with a three-level early warning decision tree and a blockchain evidence storage system. When the monitored data triggers the threshold, the entire process from early warning generation to emergency repair work order push is completed within 10 seconds, forming an intelligent operation and maintenance closed loop of monitoring, analysis, decision-making and execution.

[0021] Furthermore, the perception layer is also equipped with backup sensors, including ultrasonic sensors;

[0022] The three-level monitoring network in the perception layer includes:

[0023] The first level is point monitoring, which uses sensors deployed around the pile foundation to collect real-time data on key local points of a single pile foundation and obtain microscopic data on pile displacement and vibration.

[0024] The second level is line monitoring, which uses drone lidar scanning to scan the surface of the pile foundation and the mudline of the seabed to obtain linear data of continuous changes in the pile foundation, connecting single-point and regional monitoring;

[0025] The third level is area monitoring, which relies on In-SAR satellite inversion technology to obtain the pile foundation displacement trend at the regional scale of the wind farm, and realize large-scale, macroscopic displacement change monitoring.

[0026] Furthermore, the input parameters for the platform layer include pile foundation physical quantities, soil physical quantities, scour depth, or number of load cycles.

[0027] Furthermore, within the platform layer, a py curve correction module is embedded in the prediction model for the horizontal bearing capacity of large-diameter steel piles for wind turbines.

[0028] Furthermore, in the application layer, the finite element model is constructed by professional digital modeling software. By defining key parameters such as the nonlinear constitutive relationship of concrete and the friction characteristics of the pile-soil contact surface, in-depth simulation calculations of the bearing capacity of the pile foundation are performed. At the same time, a real-time data interface module has been developed to dynamically map real-time data to the corresponding nodes of the three-dimensional model and render it intuitively in the form of a heat map.

[0029] Therefore, the present invention employs the above-mentioned intelligent monitoring method and system for the bearing capacity of offshore wind power pile foundations, which has the following technical effects:

[0030] (1) The present invention adopts a dual-performance collaborative monitoring system of "horizontal displacement + natural frequency" to collect real-time displacement and vibration data of pile foundation, and combines In-SAR satellite inversion and UAV lidar scanning to construct a three-level monitoring network from point to surface, which solves the problems of large data deviation and lack of regional correlation in traditional methods;

[0031] (2) The present invention adopts a dual-engine architecture of physical mechanism and data-driven approach. On the one hand, it embeds a soil softening model that considers initial static shear stress and corrects the py curve to simulate the degradation of foundation stiffness under cyclic loading. On the other hand, it uses a Bayesian neural network to invert the scour depth and damping ratio, dynamically calibrates the model parameters, and realizes closed-loop optimization of theoretical calculation and measured data.

[0032] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0033] Figure 1 This is a numerical model of a pile foundation after scour in an embodiment of an intelligent monitoring method and system for the bearing capacity of offshore wind power pile foundations, wherein (a) is a schematic diagram of a three-dimensional model and (b) is a schematic diagram of two-dimensional dimensions;

[0034] Figure 2 This is a comparison of the calculation results of the horizontal displacement at the top of the pile in an embodiment of an intelligent monitoring method and system for the bearing capacity of offshore wind power pile foundations;

[0035] Figure 3 This is a schematic diagram of an intelligent monitoring system for the bearing capacity of offshore wind power pile foundations;

[0036] Figure 4 This is a sensor layout diagram in an embodiment of an intelligent monitoring method and system for the bearing capacity of offshore wind power pile foundations;

[0037] Figure 5 This is a flowchart illustrating the method for calculating the natural frequency of a monopile offshore wind turbine structure in an embodiment of an intelligent monitoring method and system for the bearing capacity of offshore wind turbine pile foundations.

[0038] Figure 6 This is a schematic diagram of an intelligent early warning platform in an embodiment of an intelligent monitoring method and system for the bearing capacity of offshore wind power pile foundations;

[0039] Figure 7 This is a flowchart illustrating the UAV lidar point cloud scanning technology in an embodiment of an intelligent monitoring method and system for the bearing capacity of offshore wind power pile foundations.

[0040] Figure 8 This is an example of an intelligent monitoring method and system for the bearing capacity of offshore wind turbine pile foundations. It shows the arrangement of acceleration sensors in a monitoring project of an offshore wind turbine in Fujian Province. (a) shows the arrangement position of the acceleration sensor platform, and (b) shows the arrangement position of the acceleration sensors on each platform.

[0041] Figure 9 This is an example of an intelligent monitoring method and system for the bearing capacity of offshore wind turbine pile foundations. It contains second-order modal frequency tracking data from a monitoring project of an offshore wind turbine in Jiangsu Province. (a) is the second-order frequency identified by the acceleration response of measuring point 1357, and (b) is the second-order frequency identified by the acceleration response of measuring point 2468.

[0042] Figure 10 This is an example of an intelligent monitoring method and system for the bearing capacity of offshore wind turbine foundations. The RMS values ​​of acceleration signals in a monitoring project of an offshore wind turbine in Jiangsu Province are measured every 15 minutes over a year. Among them, (a) is the RMS value of acceleration signals at 2468 measuring points, and (b) is the RMS value of acceleration signals at 1357 measuring points. Detailed Implementation

[0043] The present invention will be explained in more detail through the following embodiments. The purpose of disclosing the present invention is to protect all changes and modifications within the scope of the present invention. The present invention is not limited to the following embodiments.

[0044] Example 1

[0045] This invention takes a 3MW single-pillar offshore wind turbine structure as an example. It employs Cuéllar's model of soil compaction and settlement around the pile, considering long-term cyclic loading effects, to construct a three-dimensional numerical analysis model for the pile foundation that takes into account soil compaction and settlement around the pile. For example... Figure 1 As shown.

[0046] like Figure 2 As shown, the horizontal displacement at the pile top differs significantly with and without the influence of scour. Without scour, the overall fluctuation of the horizontal displacement at the pile top is relatively gentle, and the value is low. With scour, the fluctuation amplitude of the horizontal displacement at the pile top increases significantly, and the peak displacement is higher, with an overall horizontal increase in displacement. This indicates that compared to the absence of scour, scour significantly increases the horizontal displacement at the pile top, having a prominent impact on the horizontal mechanical response of the pile foundation.

[0047] Furthermore, the nonlinearity of the soil will cause uncertainty in the soil parameters. Assuming that the soil density follows a normal distribution with a standard deviation of 3%, the settlement depth around the pile will vary from 1m to 7m. The variation patterns of the overall natural frequency of the offshore wind turbine structure and the horizontal displacement of the pile top are shown in Table 1.

[0048] Table 1. Influence of soil scour depth around pile on natural frequency and horizontal displacement at pile top

[0049] ;

[0050] As shown in Table 1, the scour depth is linearly negatively correlated with the natural frequency. For every 2m increase in scour depth, the natural frequency decreases by an average of 0.003Hz. In this embodiment, this law has been embedded in the prediction process of the horizontal bearing capacity of the large-diameter steel piles of the wind turbine, and a Py curve correction module has been designed. The scour depth is positively correlated with the horizontal displacement.

[0051] In summary, scour does indeed affect the operation of offshore wind turbines. Furthermore, research into monitoring and early warning technologies for changes in pile foundation performance and frequency caused by scour revealed three major bottlenecks in current technologies: monitoring dimensions, digital twin accuracy, and structural assessment capabilities.

[0052] (1) Traditional monitoring methods face multi-dimensional technical bottlenecks in monitoring the scour of offshore wind power pile foundations. The essence is the contradiction between the harshness of the marine environment and the accuracy of monitoring requirements. Underwater visual inspection methods are constrained by environmental factors such as wave height and visibility. For example, when the wave height is >2m, the detection efficiency drops by 70%, resulting in a fundamental defect of insufficient monitoring range. This forces the technology to rely on alternative solutions such as ultrasonic waves. However, ultrasonic monitoring is prone to sensor damage due to the high flow velocity in strong hydrodynamic areas, resulting in a technical deadlock of inaccurate measurement data. Ultimately, periodic inspections are required, but because the interval is greater than 1 month, it is impossible to capture the "hourly" development rate of scour. This environmental constraint creates a causal chain of technical deficiencies and monitoring lags, specifically manifested in the following ways: while grating fiber optic sensing allows for distributed measurement, the fragility of the sensors (requiring nano-ceramic encapsulation) leads to discrete-point monitoring, making it difficult to reflect continuous changes in the scour surface; resistivity technology is affected by suspended sediment in the water, resulting in a "false positive rate exceeding 30%" of accuracy degradation in high-sediment-content sea areas; and analysis methods based on superstructure vibration suffer from assessment errors generally exceeding 15% due to difficulties in decoupling scour depth from dynamic factors. These technical shortcomings, through single-point measurement deviations causing regional data silos, create a transmission mechanism that leads to the failure of trend analysis, resulting in unplanned downtime accidents caused by pile foundation scour, exposing the structural defects of traditional methods in terms of insufficient accuracy, time lag, and difficulty in decoupling multiple parameters.

[0053] (2) Currently, there are multiple technical shortcomings in the field of offshore wind power in terms of digital twin systems and pile foundation monitoring technology. The essence is the contradiction between the compatibility between dynamic scour process and static simulation model: because the existing digital twin system relies on offline simulation data and the update cycle is greater than 24 hours, it is difficult to respond in time to the "hourly" scour development rate, resulting in a delay of at least one tidal cycle in protection decision-making, which in turn causes the pile foundation safety assessment to fall into a vicious cycle of outdated data and failure of early warning. This timeliness defect also has a superimposed effect with the accuracy bottleneck of monitoring technology. That is, although the multibeam echo sounding system can depict the seabed topography, it is affected by ocean currents and temperature gradients, such as thermoclines, which cause sound velocity deviation of more than 5%, resulting in a measurement error of 10-30cm. On the other hand, the three-dimensional imaging sonar is limited by the resolution of underwater images, and the typical defect identification accuracy is greater than 1mm, making it difficult to capture the subtle deformation in the early stage of scour. There is a transmission chain of simulation lag causing measurement distortion, leading to misjudgment of assessment. While fusing multibeam data can alleviate the aforementioned problems, it is highly dependent on data accuracy, resulting in modeling deviations exceeding 20% ​​in areas with strong ocean currents. Furthermore, it is only applicable to static terrain calculations and cannot dynamically track the impact of transient dynamic factors such as tides and waves on erosion. Its applicability to complex geological conditions or new types of foundations is severely insufficient. Operationally, it relies on specialized software and programming environments, leading to a high technical threshold for grassroots teams. At the same time, the system lacks intelligent early warning functions, only capable of post-event analysis and unable to achieve real-time risk warnings.

[0054] (3) Traditional monitoring methods mostly focus on single-point measurement of seabed soil scour depth and do not pay attention to the impact of scour on the structure. They lack the linkage analysis logic between data and structural bearing capacity, resulting in structural defects in the monitoring system. It can be seen that the existing methods have neither established a quantitative correlation model between scour depth and pile foundation bearing capacity, nor do they have real-time monitoring of the dynamic characteristics of pile foundation structures. These parameters are the core indicators for assessing the health status of structures. In actual operation, the weakening of soil constraints caused by seabed scour will directly lead to a decrease in the vertical bearing capacity of the pile foundation, excessive horizontal displacement, and ultimately manifest as the wind turbine's natural frequency shifting towards the dangerous extreme value. In addition, the existing monitoring system generally suffers from the problem of single parameter dimensions, relying only on a single sensor or physical quantity, making it difficult to form a closed loop for cross-validation of multi-source data. This results in the inability to identify potential structural risks in a timely manner and makes it difficult to meet the stringent requirements of deep-sea wind farms for the full life cycle health management of pile foundations.

[0055] To address the core issues in the operation and maintenance of offshore wind farm foundations, such as low monitoring accuracy, lack of structural bearing capacity assessment, data silos, and delayed early warning, this invention utilizes digital twin modeling of wind turbine foundations to quantify the impact mechanism of scour on frequency and bearing capacity, thereby constructing an intelligent monitoring and early warning platform for offshore wind farm foundations. Simultaneously, this platform innovatively creates a multi-collaborative intelligent monitoring system, achieving closed-loop management across the entire chain from data acquisition and real-time processing to decision output, systematically overcoming the bottlenecks of traditional operation and maintenance technologies.

[0056] like Figure 3 As shown, this invention provides an intelligent monitoring system for the bearing capacity of offshore wind power pile foundations. Based on a layered decoupled architecture, it includes a collaboratively designed perception layer, edge layer, platform layer, and application layer, achieving closed-loop management of the entire process from data acquisition to decision output, as detailed below:

[0057] At the perception layer, high-precision displacement and acceleration sensors are deployed to monitor the measured displacement and structural acceleration of the pile body at different depths when the pile foundation is subjected to the scouring effect caused by wave current. The regional displacement trend inverted by In-SAR satellite and the surface defect point cloud scanned by UAV lidar are fused simultaneously. The monitoring data is stored in the storage warehouse and collected and transferred through the wireless transmission warehouse, thus constructing a three-level monitoring network from point to surface.

[0058] The core of the three-tiered monitoring network is a collaborative network formed by three types of monitoring methods based on spatial monitoring dimensions: The first tier is "point monitoring," which uses sensors deployed around the pile foundation to collect real-time data on key local points of a single pile foundation, obtaining microscopic data such as pile displacement and vibration; the second tier is "line monitoring," which uses UAV lidar scanning to scan the pile foundation surface and seabed mudline, obtaining linear data such as pile foundation surface defects and continuous changes in seabed erosion surface, connecting single-point and regional monitoring; the third tier is "area monitoring," which relies on In-SAR satellite inversion technology to obtain the pile foundation displacement trend at the regional scale of the wind farm, achieving large-scale, macroscopic displacement change monitoring.

[0059] At the edge layer, the original signal is preprocessed in real time by an embedded server. The natural frequency is extracted using Fourier transform. Combined with wavelet denoising and Kalman filtering algorithms, the displacement data is verified from two sources, ensuring that the deviation rate is less than 5%, and providing high-precision data support for upper-level analysis.

[0060] At the platform layer, based on a customized algorithm program that considers the impact of scour on the bearing capacity of pile foundations and calculates the frequency of wind turbine structures, the main displacement and main frequency of the pile foundation are calculated by inputting the data limit values ​​in the standard specifications and combining them with the frequency spectrum characteristic curve drawn using the natural frequency of the wind turbine structure. The safety status of the pile foundation and the ultimate bearing capacity are dynamically evaluated and predicted by inputting various parameters of the pile foundation and soil.

[0061] At the application layer, with a digital twin dashboard as the core, the finite element model is integrated with real-time data to dynamically render stress concentration areas. It is equipped with a three-level early warning decision tree and a blockchain evidence storage system. When the monitoring data triggers the threshold, the entire process from early warning generation to emergency repair work order push is completed within 10 seconds, forming an intelligent operation and maintenance closed loop of monitoring, analysis, decision-making and execution.

[0062] Example 2

[0063] This invention provides an intelligent monitoring method for the bearing capacity of offshore wind turbine pile foundations, specifically including:

[0064] S1, such as Figure 4 As shown, a magnetostrictive displacement sensor and a MEMS accelerometer array are deployed to synchronously collect real-time displacement and vibration data of the pile foundation.

[0065] In marine environmental displacement monitoring of offshore wind turbine foundations, magnetostrictive displacement sensors with corrosion resistance and long-term stability are preferred for monitoring foundation displacement. These sensors support pre-embedding during the pouring stage or later installation by underwater robotic arms, and cable connections are achieved via wet-plug connectors. The data acquisition system is equipped with dual-redundant timing chips (1ns resolution), dynamically adjusting the sampling frequency as needed: 10Hz low-frequency sampling for energy saving in normal sea conditions, and 100Hz high-frequency sampling to capture transient displacements during typhoons. The edge-layer embedded acquisition unit has a built-in temperature compensation module, which, combined with the sensor, corrects the waveguide rod expansion error to ±3μm / ℃ in real time. Data is transmitted via 4G / 5G + BeiDou dual-mode, ensuring real-time communication up to 200 kilometers offshore. Signal processing employs wavelet denoising to remove wave interference and cross-validates with inclinometer data (redundant calibration is activated when the deviation exceeds 5%). Modal decomposition analysis of displacement time histories during typhoons identifies soil stiffness degradation. For deep-sea pile foundations (pile length exceeding 100 meters), a distributed array arrangement is adopted (one pre-embedded every 20 meters), and nanosecond-level time alignment is achieved through fiber optic synchronous clocks to synthesize the displacement curve of the entire pile body (error ±0.5mm); a self-sinking protective sleeve is provided to automatically cover the sensor when the scour depth exceeds 50cm, preventing marine organisms from attaching and colliding with fishing nets, providing reliable support for the monitoring of the entire life cycle of the pile foundation.

[0066] In structural acceleration monitoring, MEMS accelerometers with advantages of miniaturization, low power consumption, and integration are preferred. These sensors can be attached to the pile surface via waterproof adhesive or bolts. Their millimeter-sized, lightweight design does not affect the original structural dynamic characteristics, and they are compatible with IP68 waterproof encapsulation, enabling them to adapt to the humid environment of offshore wind turbine pile foundations. During data acquisition, a voltage signal is directly output and connected to an edge-layer embedded acquisition instrument (such as the NI 9234). Three-dimensional acceleration time histories are recorded at a sampling frequency of 100-1000Hz. An integrated temperature compensation circuit corrects for temperature drift in silicon-based materials, ensuring data stability. During data processing, the voltage signal is amplified and filtered. The natural frequency of the structure is extracted using Fourier transform, and stiffness changes are inverted using a finite element model. The displacement response is obtained by integrating the acceleration time histories twice (trapezoidal integration or integration after wavelet denoising), and cross-validated with measured values ​​from a laser displacement gauge. This sensor has high integration and can share the acquisition system with strain gauges and tilt sensors, reducing wiring complexity. With power consumption of only a few milliwatts, it supports solar or battery power, making it suitable for unmanned deep-sea scenarios. However, this sensor has drawbacks such as high low-frequency noise (susceptible to interference below 0.1Hz) and zero drift over long-term use. These can be addressed by adding a 0.5Hz high-pass filter to eliminate low-frequency disturbances, and by monthly measurement of the pile top displacement using a total station to calibrate the integrated displacement and correct the zero drift error through fitting. To address the ±50g range limitation, piezoelectric sensors can be connected in parallel under strong impact conditions to form a complementary high- and low-frequency network, ensuring data integrity under extreme loads.

[0067] S2. When offshore wind turbines are in operation, a dynamic feedback cycle occurs where, as the number of cyclic loadings increases, the anti-scour capacity of the soil decreases, the scour depth increases, the foundation stiffness decreases, and ultimately the natural vibration frequency decreases. The cyclic load promotes scour development by damaging the soil structure and intensifying water flow erosion, while scour affects the frequency by weakening the foundation support and changing the pile-soil interaction mechanism. The two are coupled to accelerate the deterioration of the pile foundation performance. It can be seen that the bearing capacity of the pile foundation is jointly affected by multiple parameters. Therefore, in engineering, multi-parameter joint monitoring and intelligent model warning are required for this process, and measures such as anti-scour protection and soil reinforcement are combined to cut off the feedback chain and ensure the safety of the pile foundation.

[0068] For this process, the present invention constructs a prediction model for the horizontal bearing capacity of large-diameter steel piles of wind turbines based on a dual-engine architecture of physical mechanism and data-driven. The physical mechanism engine embeds a soil softening model considering the initial static shear stress and a p-y curve correction module to construct a parametric physical benchmark; the data engine inversely calculates the scour depth and damping ratio through a Bayesian neural network and dynamically calibrates the model parameters to achieve closed-loop optimization of theory and measurement.

[0069] In the actual application process, the p-y curve correction module first imports basic parameters, including the physical parameters of the pile foundation, the physical parameters of the soil around the pile, and other environmental impact parameters, and at the same time calls relevant data boundary values to provide a reference basis for curve correction. Then, a soil softening model considering the influence of the initial static shear stress and cyclic pre-shear is embedded, and a weakening factor is introduced. The degree of soil stiffness degradation is calculated according to the number of cyclic loadings, and then the foundation stiffness parameters of the traditional p-y curve are adjusted. For example, when the scour depth increases, the p-y curve correction module will reduce the foundation stiffness coefficient within the corresponding depth range according to the actual scour data obtained by monitoring, so that the p-y curve can reflect the soil constraint weakening effect caused by scour. The p-y curve correction module outputs the adjusted foundation stiffness matrix and assembles it into the overall structural stiffness matrix of the large-diameter steel pile of the wind turbine, and substitutes it into the classical dynamic motion equation to calculate the natural vibration frequency and the horizontal displacement of the pile top under the action of cyclic load; at the same time, the p-y curve correction module will receive multi-source monitoring data such as the measured displacement from the displacement sensor and the natural vibration frequency extracted by the acceleration sensor transmitted by the sensing layer, and further calibrate the stiffness degradation coefficient, soil parameters, etc. of the p-y curve by comparing with the calculation results to ensure that the deviation between the natural vibration frequency calculated by the model and the measured value meets the requirements.

[0070] The dynamic calibration model parameters revolve around the core elements of pile-soil interaction and structural dynamic characteristics, specifically including three categories: First, soil physical and mechanical parameters, such as the elastic modulus, cohesion, and internal friction angle of the soil around the pile. These parameters change due to cumulative soil damage under cyclic loading and need to be calibrated to match the actual stress state. Second, foundation stiffness-related parameters, namely the stiffness coefficients in the corrected py curve (such as the initial foundation stiffness and stiffness degradation coefficient), need to be adjusted in conjunction with the inversion results to ensure that the simulated foundation stiffness degradation is consistent with the measured values. Third, structural dynamic parameters, mainly the pile damping ratio, which directly affects the accuracy of natural frequency calculation and needs to be dynamically corrected using the Bayesian neural network inversion results.

[0071] It should be noted that:

[0072] (1) The soil softening model for initial static shear stress mentioned does not directly calculate the natural frequency, but provides a key basis for frequency calculation. This model simulates the soil softening effect around the pile by considering the initial static shear stress generated by long-term horizontal cyclic loads of wind and waves, and corrects the py curve to reflect the degradation of foundation stiffness under cyclic load. The corrected foundation stiffness matrix is ​​assembled into the overall structural stiffness matrix and substituted into the classical dynamic motion equation to calculate the natural frequency of the pile foundation after N cycles of cyclic load. The calculation results need to be compared with the measured natural frequency to verify the accuracy of the model in simulating the degradation of foundation stiffness.

[0073] (2) The role of inverted scour depth and damping ratio: On the one hand, the inverted scour depth is used to correct the foundation stiffness calculation. Since the increase of scour depth will reduce the effective soil constraint within the pile foundation burial depth range, and thus reduce the foundation stiffness, the inverted value needs to be substituted into the soil softening model and the py curve parameters should be adjusted so that the calculated natural frequency is closer to the measured value. On the other hand, the inverted damping ratio is directly used to optimize the damping term in the dynamic motion equation, reduce the frequency calculation error, and at the same time, by comparing the correlation between the inverted scour depth, damping ratio and measured natural frequency, the model parameters are dynamically calibrated to achieve closed-loop optimization of theoretical calculation and measured frequency, and improve the frequency prediction accuracy.

[0074] Based on the above, and using classical dynamic motion equations, a method for calculating the natural frequency of a monopile offshore wind turbine structure was established by embedding an existing soft clay softening model that considers the effects of initial static shear and cyclic pre-shear. Figure 5As shown in the figure, the influence of pile diameter, burial depth, load amplitude, and number of cyclic loading cycles on the natural frequency of offshore wind turbine structures was investigated. Based on this calculation method, spectral characteristic curves can be automatically plotted based on the 1P / 3P frequency risk interval for the minimum offset 1P and maximum offset 3P of the natural frequency of different offshore wind turbine structures, providing a reference for judging the performance of wind turbine structures. When assessing the structural performance of a wind turbine, the measured natural frequency extracted from the real-time acceleration sensor data collected by the sensing layer via Fourier transform is compared with the 1P-3P range on the spectral characteristic curve. If the measured frequency stably falls near 1P with small fluctuations, it indicates good soil constraint around the piles, no significant degradation in foundation stiffness, and the wind turbine's structural performance is healthy. If the measured frequency shifts from 1P to 3P with a gradually increasing shift, such as approaching 2 / 3 or higher of 3P, it reflects that the soil around the piles may experience a decrease in stiffness due to cyclic loading, scouring, or other factors, leading to changes in the dynamic characteristics of the wind turbine structure. Attention should be paid to the trend of structural performance deterioration. If the measured frequency exceeds the 3P range, or rapidly shifts from 1P to 3P within a short period and approaches or even exceeds the upper limit, it indicates a significant weakening of pile-soil interaction, severe degradation of foundation stiffness, and the wind turbine's structural performance is at risk. Further comprehensive assessment, including displacement data, is needed to determine whether there are any structural safety hazards.

[0075] Based on the input structural performance parameters, and using multi-source data fusion technology, the bearing capacity of the pile foundation and the structural state of the wind turbine are accurately assessed through a dual-dimensional approach of pile displacement field analysis and structural dynamic characteristic identification. In the displacement analysis section, a Kalman filter algorithm is used to fuse the horizontal displacement values ​​measured by displacement sensors with those retrieved from In-SAR, providing dual verification of the displacement.

[0076] When the deviation between the horizontal displacement measured by the displacement sensor and the horizontal displacement retrieved by In-SAR is less than 5%, it is included in the evaluation, specifically including:

[0077] Based on offshore wind power design and testing specifications and physical models, the processed data is compared to generate an assessment report focusing on the structural health of the pile foundation and wind turbine, integrating multi-dimensional data and analysis conclusions. The report's displacement data section presents details after dual-source verification, including real-time displacement values ​​at different monitoring points, trend curves, and comparisons with specification limits, reflecting whether displacement exceeds limits or abnormal deformation occurs. The frequency analysis module lists natural frequency values ​​and fluctuation patterns, and analyzes the displacement amplitude and causes in conjunction with the safe range. The bearing capacity prediction section clearly defines the calculated ultimate bearing capacity, safety factor, and its matching degree with design requirements, quantifying the bearing capacity. The report also interprets data in accordance with industry health assessment standards, using displacement-frequency correlation analysis to determine whether foundation stiffness has degraded, and assessing whether the bearing capacity has decreased and whether it can meet the safe operation requirements under complex loads based on bearing capacity prediction, providing a reference for operation and maintenance decisions.

[0078] If the threshold is exceeded, the sensor array health diagnosis process is automatically triggered, and an early warning matrix for inelastic deformation of the pile body is generated simultaneously. For valid data, the system matches the measured displacement peak with the valid displacement data generated by the distributed computing cluster program. When the measured value reaches the displacement limit threshold specified in the standard, an early warning data packet is immediately pushed to the intelligent platform, along with real-time displacement data and a sensor array topology diagram. In the frequency analysis section, the system receives the natural frequency monitored by the accelerometer and processed. By comparing it with the spectral characteristic curve constructed based on the natural frequency of the wind turbine structure, a dynamic threshold calibration algorithm is used to determine the frequency offset. If the measured frequency falls within the specified early warning interval or reaches the maximum value, a diagnostic report containing a spectral waterfall plot and a damping ratio analysis table is generated and pushed to the intelligent early warning platform. Based on this, the system relies on the pile-soil interaction parameter library in the standard document, combines the displacement and frequency coupling analysis results, and uses a neural network model to perform time-series prediction of the ultimate bearing capacity of the pile foundation, automatically generating an operation and maintenance report with a digital twin visualization interface.

[0079] S3, such as Figure 6 As shown, the intelligent early warning platform uses 3D digital twin technology as its core carrier to construct a fully visualized operation and maintenance decision-making center. When it receives an early warning data packet transmitted from the platform layer, which contains multi-dimensional abnormal signals such as displacement exceeding limits and frequency deviation, it first dynamically analyzes the real-time damage status of the pile foundation through digital twin modeling, and simultaneously drives the safety factor dashboard to issue dynamic red, yellow, and blue early warnings. Since a rapid response needs to be triggered when the safety factor reaches a threshold, when the red early warning condition is met, the platform will automatically retrieve the entire life cycle monitoring data chain of the pile foundation for correlation analysis.

[0080] The early warning mechanism is based on a three-level trigger logic constructed according to specifications. In the first-level early warning, if the displacement deviation is between 5% and 10% or the frequency deviation is 8% of the design value, the In-SAR regional displacement data and the frequency monitoring value of the MEMS accelerometer are integrated to trigger a sensor health diagnosis report and mark the seabed scouring and frequency drift trends, providing a basis for early performance assessment. When the displacement exceeds the limit by 10%-20% or the frequency drops sharply by 10%, the second-level early warning activates the backup sensor redundancy verification, and at the same time, the UAV lidar scans the mud surface of the pile foundation to generate a three-dimensional model containing the displacement vector field and frequency characteristic curves. Combined with digital model simulation, the stiffness degradation trend is predicted. If the displacement exceeds the limit by 20% or the frequency drops sharply by 15%, the third-level early warning activates the emergency plan library in seconds. Based on the high-frequency data of the accelerometer and the lidar model, a repair work order package containing underwater emergency maintenance plan and spare parts list is generated. The system completes the response closed loop for the dual parameter anomalies of displacement and frequency within 10 seconds and automatically matches historical fault cases to generate handling procedures.

[0081] Furthermore, the maintenance work order system relies on blockchain smart contract technology to achieve full-process evidence storage from early warning reception to handling. Work orders are automatically associated with real-time sensor data streams, reference to standard clauses, and expert consultation opinions. The system recommends the optimal maintenance path through knowledge graph technology and simultaneously pushes task instructions with digital signatures to relevant units. The platform supports multi-terminal interaction, and maintenance personnel can intuitively view hidden defects such as pile foundation scour surfaces and crack development through a digital twin interface. They can also obtain maintenance process references by combining on-site operation guidance documents, thus forming an intelligent maintenance closed loop from early warning to diagnosis, handling, and source tracing.

[0082] Example 3

[0083] This invention provides an intelligent monitoring system for the bearing capacity of offshore wind power pile foundations. It constructs a digital twin-based visualization hub, integrates the pile-soil finite element model with real-time sensor data, dynamically renders stress concentration areas and supports three-dimensional cross-sectional viewing, so as to achieve full-chain intelligence from monitoring to decision-making.

[0084] This invention uses a self-coded prediction model for the horizontal bearing capacity of large-diameter steel piles for wind turbines to correct the foundation stiffness of large-diameter piles, thereby deriving a new corrected Py curve and calculating the accurate horizontal displacement at the pile top of large-diameter steel piles with a diameter greater than 2m, providing a reference for engineering design. In practical applications, the calculation benchmark is determined by inputting pile foundation physical quantities (i.e., pile diameter, wall thickness, and embedment depth), soil physical quantities (i.e., internal friction angle of sand, effective unit weight), and other influencing factors (i.e., scour depth, number of load cycles) and industry standards. Then, based on the dynamic motion equations, a soft clay softening model is embedded to calculate the foundation stiffness matrix, and the Py curve is corrected using a weakening factor to simulate stiffness degradation. For different offshore wind turbines, the system automatically calculates and generates corresponding natural frequencies and horizontal displacements for data evaluation.

[0085] Combining In-SAR technology, microwave signals are transmitted by satellites and reflected signals are received. A phase difference map is generated through interferometric processing. Then, operations such as removing the flat-ground effect and phase unwrapping are performed to convert the radar line-of-sight displacement to a horizontal displacement. Simultaneously, satellites with a 12-day revisit cycle, such as Sentinel-1, are used to acquire high-frequency imagery. Permanent scatterer technology is used to pinpoint stable scattering points at the top of the pile foundation to eliminate interference from ocean waves and the atmosphere. In pile foundation displacement inversion, not only is the regional displacement trend of multiple pile foundations obtained through macro-monitoring covering tens of square kilometers of sea area, but also fine-grained single-point analysis is performed using the differences in scattering characteristics between the pile foundation and the surrounding soil. A time-series trend model is constructed using several years of image sequences to identify abnormal displacement rate increases. When verifying and applying data, if the deviation rate between the In-SAR inverted displacement and the field sensor data is less than 5%, it is included in the evaluation model; when the displacement deviation rate reaches 5%-10%, a first-level early warning is triggered and a scour trend report is pushed out, providing macroscopic displacement boundary conditions for the digital twin model to correct the soil stiffness parameters. At the same time, regional protection plans are formulated in combination with UAV lidar data, thereby realizing the full-chain application from data acquisition to operation and maintenance decision-making.

[0086] The drone-borne LiDAR data is acquired by a LiDAR module on the drone, enabling all-weather, non-contact scanning of offshore pile foundations. The LiDAR uses a laser beam to achieve a 360° surround scan, combined with an inertial navigation system and centimeter-level positioning technology, to collect 240,000 points per second of 3D coordinates and reflection intensity data of the pile foundation surface. A single flight can cover multiple pile foundations within a radius of 50 meters. After scanning, the raw point cloud data is denoised, stitched, and geographically converted to form a high-precision 3D point cloud model. The point cloud is then compared with the design model at the millimeter level to automatically identify damage features such as cracks, corrosion, and deformation in the pile body (with an accuracy of up to 0.5 mm).

[0087] In addition, the system has a built-in AI algorithm that can analyze seabed scour depth through point cloud density changes, simultaneously generating a damage assessment report with heat map markers, and linking it with an intelligent early warning platform. When indicators such as scour depth and structural deformation exceed the specified thresholds, it automatically triggers a level two or three early warning mechanism, pushing out emergency repair work orders containing 3D visualization models, such as... Figure 7 As shown, compared with traditional manual inspection methods, efficiency is improved by 80%, and the data repetition measurement error is <1%, providing high-timeliness and high-precision digital support for the intelligent operation and maintenance of offshore wind power pile foundations.

[0088] Example 4

[0089] This invention provides an intelligent monitoring method and system for the bearing capacity of offshore wind turbine foundation piles, applied to a monitoring project for an offshore wind turbine in Fujian Province. It employs an 8-channel horizontal acceleration sensor array, symmetrically arranged along four platforms at 0.5m, 10m, 47m, and 71m of the tower. Figure 8As shown in (a), each layer has two sensors pointing towards the prevailing wind direction and the vertical wind direction respectively, forming an X / Y bidirectional vibration monitoring network, as shown in (a). Figure 8 As shown in (b). The sensor selected is a PCB 352C65 piezoelectric accelerometer (range ±50g, resolution 10). -6 g), paired with the NI 9234 data acquisition module, captures structural vibration signals in real time at a sampling frequency of 20Hz, and transmits them to the server via a 4G network. A set of 900-second time-domain data is generated every 15 minutes, and environmental parameters such as wind speed and rotational speed are synchronously accessed from the SCADA system to form a multi-physical quantity collaborative monitoring system.

[0090] The SSI-DATA (Scattered Subspace Recognition) algorithm was used to process the acceleration time history: first, wavelet denoising was used to remove environmental noise, and then the feature system was used to extract structural modal parameters. The results showed that the first-order natural frequency of the wind turbine remained stable at around 0.33Hz, with a deviation of less than 3% from the theoretical calculation. The damping ratio increased nonlinearly with increasing wind speed; when the wind speed exceeded 12m / s, the damping ratio in the X direction increased from 0.48% to 3.2%, verifying the "wind speed-damping" coupling effect. During data processing, by comparing the RMS values ​​(root mean square) of sensors at different heights, it was found that the vibration amplitude at the upper part of the tower (71 meters) was 40% higher than that at the lower part, which is consistent with the finite element simulation results.

[0091] The system employs a three-tiered early warning logic: when the natural frequency deviates from the design value by 15% (e.g., <0.28Hz) or the damping ratio drops sharply by 20%, a level one early warning is triggered, and a spectrum analysis report is pushed out; if the measured displacement exceeds the limit (>50mm) and is accompanied by a sharp drop in frequency, a level two early warning is activated, and redundant verification of backup sensors is initiated; during typhoons (e.g., during the passage of Typhoon Tapah in September 2019), when the peak acceleration exceeds 1g, a level three early warning is automatically activated, generating a repair work order containing a three-dimensional vibration cloud map. In practical applications, this mechanism successfully detected the frequency drop phenomenon caused by pile foundation scour. Specifically, when a pile foundation experienced seabed scour to a depth of 2.5 meters, its natural frequency dropped from 0.33Hz to 0.29Hz. The system issued a structural stiffness degradation warning 48 hours in advance, and divers confirmed that the scour pit morphology matched the warning result.

[0092] In practical applications, the advantages of this invention system include: First, high accuracy. The error between the pile foundation vibration frequency measured by the sensor array and the ANSYS simulation value is less than 2%, and the displacement inversion accuracy reaches 0.5mm, meeting the specifications. Second, real-time advantage. The delay from data acquisition to early warning generation is less than 10 seconds, which is 20 times more efficient than traditional manual inspection and reduces the annual operation and maintenance cost of a single wind turbine by 1.8 million yuan. Third, multi-parameter coordination. The frequency and damping joint evaluation model constructed by integrating acceleration, wind speed, and rotational speed data enables the pile foundation health status identification accuracy to reach 92%, successfully avoiding three unplanned shutdown accidents caused by scouring.

[0093] Example 5

[0094] This invention provides an intelligent monitoring method and system for the bearing capacity of offshore wind turbine foundation piles, applied to a monitoring project for offshore wind turbines in Jiangsu Province. A "time-frequency domain joint analysis" framework was developed, fusing eight-channel acceleration data using Kalman filtering in the time domain and extracting instantaneous frequency features using Hilbert-Huang transform in the frequency domain. The method was applied to an abnormal vibration event on September 11, 2019 (peak acceleration 1.16 m / s²). 2 The system, through modal decomposition, discovered that the second-order frequency decreased from 1.6Hz to 1.4Hz. Combined with the UAV lidar point cloud scanning results, the system located a 0.8m deep pile foundation at the mud surface. 2 The corroded region was examined to verify the mapping relationship between frequency shift and corrosion damage, such as... Figure 9 As shown.

[0095] In addition, an RMS value threshold analysis (<0.01m / s) was introduced. 2 ),like Figure 10 As shown, after removing strong wind interference data, the statistical accuracy of the first-order damping ratio improved to 0.46% (median), which is 13% higher than that of the Fujian project.

[0096] Furthermore, the project employed an early warning model combining digital twins and machine learning. First, a pile-soil coupled finite element model was established using Abaqus, and UAV lidar point cloud data was imported to generate a three-dimensional damage map. Then, an LSTM neural network was used to train a prediction model for scour depth and frequency shift. An early warning was triggered when the deviation between the measured frequency and the predicted value exceeded 8%. During a typhoon in July 2020, the system predicted a horizontal displacement of 38 mm (design limit 50 mm) in the pile foundation based on acceleration time history. Simultaneous laser scanning showed a seabed scour depth of 2.1 meters, allowing underwater grouting reinforcement to be initiated three days in advance, thus avoiding the risk of structural instability.

[0097] In this engineering application, the advantages of the system of this invention include: First, breakthrough in accuracy: the laser point cloud scanning achieves an identification accuracy of 0.5mm for defects on the pile foundation surface, and the deviation between the acceleration integral displacement and the inclinometer measured value is <3%, realizing millimeter-level damage monitoring; Second, adaptability to extreme working conditions: during the passage of Typhoon Tapah, the system continuously collected effective data, measuring a maximum acceleration of 0.85g, corresponding to a sudden increase in the first-order damping ratio to 4.2%, providing a measured basis for the typhoon-resistant design of the pile foundation; Third, realization of full life cycle management: through 4G+BeiDou dual-mode transmission, real-time data backhaul is achieved in the sea area 200 kilometers offshore, and combined with blockchain evidence storage technology, the traceability of monitoring data is improved to 99.9%, and the efficiency of operation and maintenance decision-making is improved by 40%.

[0098] Therefore, the present invention adopts the above-mentioned intelligent monitoring method and system for the bearing capacity of offshore wind power pile foundations. When applied to actual engineering scenarios, it achieves improved accuracy and reduced cost. At the same time, it integrates fiber optic sensing to form a "point, line, and surface" monitoring network, and combines 5G / BeiDou to achieve low-latency transmission in deep sea areas, reducing the prediction time to the second level.

[0099] Finally, it should be noted that the above 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 preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. An intelligent monitoring method for offshore wind pile foundation bearing performance, characterized in that, The method comprises the following steps: S1, collecting real-time displacement and vibration data of the offshore wind pile foundation by displacement sensors and acceleration sensors, and performing signal processing on the collected data to extract the measured natural vibration frequency and horizontal displacement of the offshore wind pile foundation; S2, based on the dual-engine architecture of physical mechanism and data driving, a wind turbine large-diameter steel pile horizontal bearing capacity prediction model is constructed, and a multi-source data fusion technology is used to analyze the pile foundation horizontal displacement field and identify the structural dynamic characteristics to evaluate or warn the pile foundation bearing capacity and the wind turbine structure state; In the dual-engine architecture based on physical mechanism and data driving, the wind turbine large-diameter steel pile horizontal bearing capacity prediction model is constructed, specifically: the physical mechanism engine is embedded with a soil softening model considering initial static shear stress and a p-y curve correction module to construct a parameterized physical benchmark, the soil softening model based on initial static shear stress is used to correct the p-y curve, the foundation stiffness degradation under cyclic load is simulated, and the pile foundation natural vibration frequency after N times of cyclic load is calculated through the classical dynamic motion equation to verify the measured natural vibration frequency; the data driving engine inverses the scour depth and damping ratio through the Bayesian neural network to dynamically calibrate the related parameters involved in the model and optimize the pile foundation natural vibration frequency calculated by the model; Through the pile foundation horizontal displacement field analysis and the structural dynamic characteristic identification, the pile foundation bearing capacity and the wind turbine structure state are evaluated or warned, specifically: in the displacement analysis part, the Kalman filtering algorithm is used to fuse the measured pile foundation horizontal displacement and the pile foundation horizontal displacement inversely calculated by the In-SAR technology for double verification of displacement; in the frequency analysis part, the natural vibration frequency monitored by the acceleration sensor and processed is received, compared with the frequency spectrum characteristic curve constructed based on the inherent frequency of the wind turbine structure, and the dynamic threshold calibration algorithm is used to judge the frequency deviation; relying on the pile-soil interaction parameter library in the specification file, combined with the displacement and frequency coupling analysis results, the neural network model is used to time series predict the ultimate bearing capacity of the pile foundation; The evaluation process includes integrating different dimension data and analysis conclusions to generate an evaluation report focusing on the health of the pile foundation and the wind turbine structure; the warning adopts a three-level threshold warning mechanism.

2. The method according to claim 1, characterized in that, In S2, the In-SAR technology is used to generate a phase difference map by interference processing of the received signal, and then convert the radar line-of-sight displacement into horizontal displacement to inversely calculate the pile foundation horizontal displacement for double verification of displacement.

3. The method according to claim 2, characterized in that, The pile foundation horizontal displacement inversion includes obtaining the regional displacement trend of multiple pile foundations through regional macro monitoring, and performing single-point fine analysis using the scattering characteristic difference between the pile foundation and the surrounding soil; it also includes constructing a time series trend model based on years of image sequences to identify displacement acceleration anomalies.

4. The method according to claim 3, characterized in that, In S2, when the displacement deviation is less than 5%, it is included in the evaluation; otherwise, the three-level threshold warning mechanism is triggered: When the displacement deviation is between 5% and 10% or the natural vibration frequency deviation is 8% of the design value, the first-level warning is triggered, the regional displacement trend and the measured pile foundation horizontal displacement are integrated to generate a sensor health diagnosis report, and a pile foundation inelastic deformation warning matrix is also generated. When the displacement exceeds 10%-20% or the natural frequency drops by 10%, a secondary warning is triggered, the backup sensor redundancy verification is started, and the unmanned aerial vehicle laser radar is linked to scan the pile foundation mud surface to generate a three-dimensional model containing displacement vector field and frequency characteristic curve, combined with digital model simulation to predict the stiffness degradation trend; When the displacement exceeds 20% or the natural frequency drops by 15%, a tertiary warning is triggered, based on acceleration sensor high-frequency data and unmanned aerial vehicle laser radar data, the emergency pre-library is activated to generate an emergency repair work order package containing underwater emergency repair schemes and spare parts list.

5. An intelligent monitoring system for offshore wind pile foundation bearing performance, used to implement the intelligent monitoring method for offshore wind pile foundation bearing performance according to any one of claims 1-4, characterized in that, It includes a perception layer, an edge layer, a platform layer, and an application layer for collaborative design, forming a full-process closed-loop management from data acquisition to decision output; The perception layer is deployed with displacement sensors and acceleration sensors to monitor pile foundations of different depths in real time and store monitoring data in storage bins for data collection and transfer through wireless transmission bins. In addition, it integrates In-SAR satellite and unmanned aerial vehicle laser radar monitoring to build a three-level monitoring network from point to surface, with different monitoring methods at each level to achieve point-to-surface and micro-to-macro pile foundation data monitoring. The edge layer uses embedded servers to perform real-time preprocessing of raw signals, uses Fourier transform to extract natural frequencies, and uses wavelet denoising and Kalman filtering algorithms for double-source verification of displacement data. The platform layer is based on the wind turbine large-diameter steel pile horizontal bearing capacity prediction model, and uses the frequency spectrum characteristic curve drawn based on the input data limit values in the standard specification and the inherent frequency of the wind turbine structure to dynamically evaluate the safety state of the pile foundation and predict the ultimate bearing capacity. The application layer is equipped with an intelligent warning platform centered on a digital twin board, which integrates finite element models with real-time data to dynamically render stress concentration areas, and is equipped with a three-level warning decision tree and a blockchain storage system. When the monitoring data triggers the threshold, the full-process response from warning generation to repair work order pushing is completed within 10 seconds, forming an intelligent operation and maintenance closed loop of monitoring, analysis, decision-making, and execution.

6. The intelligent monitoring system for bearing performance of offshore wind power pile foundation according to claim 5, characterized in that, The perception layer is also deployed with backup sensors, including ultrasonic sensors. The three-level monitoring network in the perception layer includes: The first level is point monitoring, which realizes real-time data collection of local key points of a single pile foundation by deploying sensors around the pile foundation, and obtains micro data of the pile body; The second level is line monitoring, which scans the surface of the pile foundation and the sea bed mud surface line through the unmanned aerial vehicle laser radar, obtains the linear data of the continuous change of the pile foundation, and connects single-point and area monitoring; The third level is surface monitoring, which relies on In-SAR satellite inversion technology to obtain the displacement trend of the pile foundation at the regional scale of the wind farm, and realizes large-scale and macro displacement change monitoring.

7. The intelligent monitoring system for bearing performance of offshore wind power pile foundation according to claim 5, characterized in that, The input parameters of the platform layer include pile foundation physical quantities, soil physical quantities, scouring depth, or load cycle number.

8. The intelligent monitoring system for bearing performance of offshore wind power pile foundation according to claim 5, characterized in that, In the platform layer, the p-y curve correction module is embedded in the wind turbine large-diameter steel pile horizontal bearing capacity prediction model, which dynamically optimizes the simulation accuracy of the traditional p-y curve for pile-soil interaction by combining the complex environmental factors of the sea and the measured data. The p-y curve correction module introduces a weakening factor to calculate the soil stiffness degradation degree according to the number of cyclic loads, and then adjusts the foundation stiffness parameters of the traditional p-y curve.

9. The intelligent monitoring system for bearing performance of offshore wind power pile foundation according to claim 5, characterized in that, In the application layer, the finite element model is constructed by professional digital modeling software. By defining the key parameters of the nonlinear constitutive relation of concrete and the friction characteristics of the pile-soil contact surface, the depth simulation calculation of the bearing capacity of the pile foundation is carried out. At the same time, a real-time data interface module is developed to dynamically map the real-time data to the corresponding nodes of the three-dimensional model and intuitively render them in the form of a heat map.

Images