Simulation Analysis Method of Steel Pile Hydraulic Cylinder Stress and Steel Pile Center of Gravity Change Model

Abstract

The present application relates to the technical field of steel pile engineering simulation analysis, and discloses a simulation analysis method for force of a steel pile oil cylinder and a model of change of the center of gravity of the steel pile, comprising the following steps: obtaining oil cylinder pressure, steel pile displacement trajectory data and geological parameters and cleaning; extracting multi-dimensional features to generate dynamic load distribution parameters and constructing a coupling relationship model; jointly verifying the oil cylinder pressure and the dynamic displacement of the steel pile, identifying the spatial correspondence of pressure gradient change and center of gravity deviation, and quantifying abnormal correlation; generating a system stability index to determine whether to trigger control parameter correction; if correction is required, a pressure compensation scheme is automatically generated. Through multi-dimensional data processing, coupling modeling and dynamic verification, the method realizes accurate analysis and intelligent control of the force of the oil cylinder and the change of the center of gravity in steel pile operation, improves the system stability and construction safety, and is suitable for steel pile construction scenes such as ocean engineering.

Description

Technical Field

[0001] This invention relates to the field of simulation analysis technology for steel pile engineering, specifically to a simulation analysis method for the force on the hydraulic cylinder of a steel pile and the change in the center of gravity of the steel pile. Background Technology

[0002] In marine engineering, port construction, and other fields, steel pile construction is a critical and complex foundation engineering operation. The accurate driving and stable bearing capacity of steel piles directly affect the safety, reliability, and economy of the entire project. During steel pile operation, the hydraulic cylinder, as the core actuator driving the steel pile to sink, has a complex dynamic coupling relationship with the stress state of the cylinder and the change of the steel pile's center of gravity. This relationship is significantly affected by various factors, including geological environmental parameters (such as soil shear strength and seawater impact force distribution) and the steel pile's displacement trajectory (such as changes in inclination angle and vertical displacement deviation).

[0003] In traditional steel pile construction, the analysis of hydraulic cylinder stress and changes in the steel pile's center of gravity often relies on relatively simple monitoring methods and empirical judgment. For example, simply monitoring the hydraulic cylinder pressure value to roughly assess the settlement state of the steel pile, or manually observing changes in the pile's inclination angle to judge its stability, lacks comprehensive analysis of multi-dimensional data and in-depth research on dynamic coupling relationships. This makes it difficult to accurately capture the real-time correlation between hydraulic cylinder pressure fluctuations and the steel pile's center of gravity shift. Consequently, under complex geological conditions or harsh sea conditions, abnormal mechanical correlations during construction cannot be detected in time, easily leading to engineering accidents such as steel pile tilting, shifting, or even breakage, seriously affecting construction progress and project quality.

[0004] Furthermore, as marine engineering expands into deeper waters and more complex geological conditions, the environmental challenges faced by steel pile construction are becoming increasingly severe. Factors such as the dynamic changes in seawater impact and the uneven distribution of soil properties cause the load distribution during steel pile operation to exhibit highly nonlinear and time-varying characteristics. Traditional analysis methods, unable to effectively handle multi-source heterogeneous data (such as cylinder pressure data, displacement trajectory data, and geological environmental parameters), struggle to construct accurate coupling relationship models. This results in a lack of scientific rigor and reliability in assessing system stability, failing to provide effective theoretical basis for correcting control parameters during construction, and ultimately leading to insufficient risk control capabilities during construction.

[0005] Meanwhile, with the trend towards automation and intelligentization in steel pile construction, there is an urgent need for a method to perform real-time simulation analysis of the hydraulic cylinder stress and changes in the steel pile's center of gravity, in order to meet the requirements for precise control and risk warning in engineering construction. Most existing simulation analysis methods are based on simplified mechanical models, neglecting the combined influence of various factors during actual operation, resulting in significant deviations between simulation results and actual working conditions, and failing to provide reliable guidance for engineering practice. Therefore, how to integrate multi-dimensional data, construct an accurate coupling relationship model, and achieve dynamic monitoring, anomaly identification, and system stability assessment of the hydraulic cylinder stress and changes in the steel pile's center of gravity has become a key technical challenge that urgently needs to be solved in the field of steel pile engineering. Summary of the Invention

[0006] The purpose of this invention is to provide a simulation analysis method for the stress on the hydraulic cylinder of a steel pile and the change in the center of gravity of the steel pile, so as to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a simulation analysis method for the force on a steel pile hydraulic cylinder and the change in the center of gravity of the steel pile, the method comprising:

[0008] S1: Acquire hydraulic cylinder pressure data, steel pile displacement trajectory data, and geological environment parameters during the steel pile operation process, and perform data cleaning on the hydraulic cylinder pressure data, displacement trajectory data, and geological parameters;

[0009] S2: Perform multi-dimensional feature extraction on the cleaned data to generate dynamic load distribution parameters, and construct a coupling relationship model between the hydraulic cylinder force and the change of the steel pile center of gravity based on the load distribution parameters;

[0010] S3: Based on the coupling relationship model, the real-time dynamics of cylinder pressure and steel pile displacement are jointly verified. The spatial correspondence between cylinder pressure gradient change and steel pile center of gravity offset is identified through multi-axis motion correlation technology. The abnormal correlation between cylinder pressure fluctuation and center of gravity offset is quantitatively identified.

[0011] S4: Generate system stability indices based on the verification results of the coupling relationship model, and determine whether to trigger control parameter correction based on the system stability indices.

[0012] Preferably, the cylinder pressure data includes the pressure pulsation value of the main cylinder and the peak pressure of the auxiliary cylinder; the displacement trajectory data includes the time sequence of the steel pile inclination angle change and the vertical displacement deviation; the geological environment parameters include soil shear strength parameters and seawater impact force distribution; the cylinder pressure data, displacement trajectory data and geological parameters are cleaned, including noise filtering, data alignment, anomaly removal and time-domain synchronization processing; after cleaning the cylinder pressure data, a pressure feature sequence is generated, and after cleaning the displacement trajectory data, a displacement feature sequence is generated; the pressure feature sequences and displacement feature sequences of different dimensions are dynamically matched to form a multi-source load feature set.

[0013] Preferably, step S2 includes the following steps:

[0014] S201: Extract load distribution characteristic parameters from the pressure characteristic sequence, displacement characteristic sequence and geological environment parameters after cleaning. The load distribution characteristic parameters include cylinder pressure gradient, center of gravity offset rate and soil resistance coefficient. Combine each load distribution characteristic parameter according to a preset correlation degree to generate dynamic load distribution parameters.

[0015] S202: Set an adaptive threshold for the dynamic load distribution parameters in the multi-source load feature set, filter the feature parameters that meet the threshold range to form an initial load feature set, and exclude the feature parameters that exceed the threshold range.

[0016] S203: Perform cross-condition correlation analysis on each characteristic parameter in the initial load characteristic set, extract the deviation of the characteristic parameter in different conditions of the same operation stage, calculate the standard deviation of the deviation and mark it as the inter-condition disturbance parameter.

[0017] S204: Compare the disturbance parameters between operating conditions with the preset stability threshold, select the characteristic parameters that exceed the stability threshold and add them to the coupling relationship model, and supplement the abrupt change characteristics of seawater impact force into the coupling relationship model according to the geological environment parameters.

[0018] Preferably, in S3, the quantitative identification of the abnormal correlation between cylinder pressure fluctuation and center of gravity shift includes the following steps:

[0019] S301: Real-time monitoring of the cylinder pressure change rate and the acceleration of the steel pile center of gravity offset, calculating the difference in dynamic response between the two, and determining an abnormal mechanical correlation if the difference in dynamic response exceeds the preset safety range.

[0020] S302: The number of abnormal mechanical correlations within the preset operation cycle is recorded as N, and the fluctuation amplitude of soil resistance in the geological environment parameters is recorded as S.

[0021] S303: Based on the interaction between the number of abnormal mechanical correlations N and the fluctuation amplitude of soil resistance S, a dynamic correlation function is used to generate a system stability index. The multiplicative and difference components of N and S are used to generate a comprehensive evaluation value through composite calculation. If the system stability index exceeds the preset failure threshold, a control parameter correction command is triggered.

[0022] Preferably, in S4, the verification results of the coupling relationship model in the historical data of the same type of steel pile operation are extracted, and the deviation value of each model is accumulated over time to generate a system stability index. If the system stability index continuously exceeds the preset trigger threshold, it is determined that the control parameters need to be corrected.

[0023] Preferably, in S202, the method for setting the adaptive threshold includes: statistically analyzing the dynamic distribution range of each feature parameter based on the historical job database, calculating the fluctuation range of the feature parameter using a sliding window algorithm, using the upper and lower limits of the fluctuation range as the benchmark values ​​of the adaptive threshold, and dynamically correcting the threshold range according to the feature parameter distribution of the real-time job data.

[0024] Preferably, in S303, the specific implementation of the dynamic correlation function is as follows: a composite function is established with the number of abnormal mechanical correlations N as the input variable and the soil resistance fluctuation amplitude S as the correction variable. A prediction model of the system stability index is obtained by training through a finite element simulation algorithm. The prediction model adopts a multibody dynamics structure, with the input layer including the standardized values ​​of N and S, and the output layer being the system stability index.

[0025] Preferably, the method further includes S5: when it is determined that control parameter correction needs to be triggered, an oil cylinder pressure compensation scheme is automatically generated. The pressure compensation scheme includes the main oil cylinder pressure adjustment gradient, the auxiliary oil cylinder pressure limit value and the center of gravity offset compensation coefficient, and the execution priority of the compensation scheme is graded according to the degree of deviation of the system stability index.

[0026] Preferably, in S5, the method for generating the pressure compensation scheme includes: constructing a multi-constraint optimization function, taking the minimization of cylinder energy consumption, the maximization of center of gravity offset suppression, and the optimization of system response speed as constraints, using a particle swarm optimization algorithm to perform multi-parameter optimization calculations, and outputting a non-dominated solution set as a candidate pressure compensation scheme.

[0027] Preferably, the multi-constraint optimization function verifies the robustness of the solution set through Monte Carlo simulation, and adopts a dynamic weight adjustment mechanism to balance the conflict relationship between various constraints, ensuring that the candidate pressure compensation scheme meets the operational requirements under different sea conditions.

[0028] Compared with the prior art, the beneficial effects of the present invention are:

[0029] In the data processing and feature extraction stages, data cleaning operations such as noise filtering, data alignment, anomaly removal, and time-domain synchronization were performed on the hydraulic cylinder pressure data, steel pile displacement trajectory data, and geological environment parameters, effectively improving the quality and reliability of the raw data. Simultaneously, by extracting multi-dimensional load distribution characteristic parameters such as hydraulic cylinder pressure gradient, center of gravity offset rate, and soil resistance coefficient, and performing dynamic matching and cross-condition correlation analysis, the interaction relationships between various factors during steel pile operation can be comprehensively captured, laying a solid data foundation for constructing an accurate coupling relationship model. This multi-dimensional data processing and feature extraction method overcomes the limitations of traditional single-data monitoring, enabling a more comprehensive and in-depth revelation of the mechanical laws governing the steel pile operation process.

[0030] In the construction of the coupling relationship model, effective feature parameters are screened by setting adaptive thresholds, and geological environmental parameters are combined to supplement the abrupt changes in seawater impact force, enabling the model to dynamically adapt to changes in different operating conditions and environmental conditions. The adaptive threshold setting is based on a historical operating database and a sliding window algorithm, which can dynamically adjust the threshold range according to real-time operating data, ensuring that the model can always accurately capture key feature parameters, thus improving the model's robustness and adaptability. Furthermore, cross-condition correlation analysis is used to extract disturbance parameters between operating conditions, and these parameters are compared with preset stability thresholds to effectively eliminate interference factors, further improving the model's accuracy and reliability. This data-driven and dynamically adjusted model construction method significantly improves the ability to model complex operating conditions, enabling the model to more realistically reflect the mechanical coupling relationships in actual operating processes.

[0031] In terms of model validation and anomaly identification, by real-time monitoring of the hydraulic cylinder pressure change rate and the acceleration of the steel pile's center of gravity offset, and calculating the difference in their dynamic responses, abnormal mechanical correlations can be detected in a timely manner. Combining the number of abnormal mechanical correlations with the amplitude of soil resistance fluctuations, a dynamic correlation function is used to generate a system stability index, achieving a quantitative assessment of system stability. The dynamic correlation function, based on a prediction model trained using a finite element simulation algorithm and employing a multibody dynamics structure, accurately reflects the combined impact of the number of abnormal mechanical correlations and the amplitude of soil resistance fluctuations on system stability. This quantitative identification method overcomes the subjectivity and ambiguity of traditional experience-based judgments, providing a scientific and accurate basis for risk warning during construction, timely detection of potential safety hazards, and prevention of engineering accidents.

[0032] In the system stability assessment and control parameter correction phase, the verification results of the coupling relationship model are extracted from historical data of similar steel pile operations. The model deviation values ​​are accumulated over time to generate a system stability index, enabling dynamic assessment of the system's stability trends. When the system stability index exceeds a preset threshold, a control parameter correction command is triggered, and a hydraulic cylinder pressure compensation scheme is automatically generated. This pressure compensation scheme, based on multi-constraint optimization functions and particle swarm optimization algorithms, uses minimizing hydraulic cylinder energy consumption, maximizing center of gravity offset suppression, and optimizing system response speed as constraints. It achieves multi-parameter optimization calculations and provides the optimal pressure compensation scheme based on the degree of deviation from the system stability index. Simultaneously, Monte Carlo simulation is used to verify the robustness of the solution set and the dynamic weight adjustment mechanism to balance conflicting constraints, ensuring that the compensation scheme effectively improves system stability and operational efficiency under different sea conditions. This closed-loop control mechanism achieves full automation and intelligence from data acquisition, model construction, verification and evaluation to control correction, significantly improving the precision control and risk response capabilities of steel pile construction, reducing manual intervention costs, and enhancing construction efficiency and project quality. Attached Figure Description

[0033] Figure 1 This is a schematic diagram illustrating the working principle of the simulation analysis method for the force on the steel pile hydraulic cylinder and the change in the center of gravity of the steel pile described in this invention.

[0034] Figure 2 A diagram illustrating the working principle of data processing and feature set formation;

[0035] Figure 3 A schematic diagram illustrating the working principle of the coupling relationship model;

[0036] Figure 4 A schematic diagram illustrating the working principle of anomaly correlation quantitative identification and system stability index generation. Detailed Implementation

[0037] 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.

[0038] Please see Figures 1-4 The present invention relates to a simulation analysis method for the stress on the hydraulic cylinder of a steel pile and the change in the center of gravity of the steel pile. The specific implementation steps are as follows:

[0039] S1: Data Acquisition and Cleaning: Acquire multi-source data during the steel pile operation, including hydraulic cylinder pressure data, steel pile displacement trajectory data, and geological environmental parameters. The hydraulic cylinder pressure data includes the pressure pulsation value of the main hydraulic cylinder and the peak pressure of the auxiliary hydraulic cylinder. The displacement trajectory data covers the temporal sequence of steel pile inclination angle changes and vertical displacement deviation. The geological environmental parameters include soil shear strength parameters and seawater impact force distribution. Data cleaning operations are performed on the above data, specifically including noise filtering, data alignment, anomaly removal, and time-domain synchronization. Random interference during signal acquisition is removed using noise filtering algorithms. Time-domain alignment of multi-source data is achieved using timestamp synchronization technology. Invalid data points caused by equipment failure or operational abnormalities are removed. Finally, pressure feature sequences and displacement feature sequences are generated, and feature sequences of different dimensions are dynamically matched to form a multi-source load feature set.

[0040] S2: Coupled Relationship Model Construction: Load distribution characteristic parameters, including cylinder pressure gradient, center of gravity offset rate, and soil resistance coefficient, are extracted from the pressure characteristic sequence, displacement characteristic sequence, and geological environment parameters after cleaning. These characteristic parameters are combined according to a preset correlation degree to generate dynamic load distribution parameters. An adaptive threshold is then set for the parameters in the multi-source load characteristic set. This threshold is based on the dynamic distribution range of each characteristic parameter statistically analyzed from a historical operation database. A sliding window algorithm is used to calculate the fluctuation range, with the upper and lower limits of the range serving as benchmark values, and dynamically adjusted according to real-time data. Characteristic parameters that meet the threshold range are selected to form an initial load characteristic set. After excluding outliers, cross-condition correlation analysis is performed on the parameters in the initial set to extract the deviation under different conditions in the same operation stage. The standard deviation of the deviation is calculated as the inter-condition disturbance parameter. The disturbance parameter is compared with a preset stability threshold, and parameters exceeding the threshold are added to the coupled relationship model. The abrupt change characteristics of seawater impact force are supplemented based on geological environment parameters to complete the model construction.

[0041] S3: Dynamic Verification and Anomaly Identification: Real-time dynamic verification of cylinder pressure and steel pile displacement is performed based on a coupling relationship model. Multi-axis motion correlation technology is used to identify the spatial correspondence between changes in cylinder pressure gradient and the offset of the steel pile's center of gravity. Specifically, the difference in dynamic response between the cylinder pressure change rate and the steel pile's center of gravity offset is calculated by real-time monitoring. If the difference exceeds a preset safety range, it is determined to be an abnormal mechanical correlation. The number of abnormal correlations within a preset operating cycle is recorded as N, and the fluctuation amplitude of soil resistance in geological environmental parameters is simultaneously acquired and recorded as S. A dynamic correlation function is used to generate a system stability index. This function uses N as the input variable and S as the correction variable. A prediction model of the multibody dynamics structure is trained using a finite element simulation algorithm. The input layer consists of standardized values ​​of N and S, and the output layer is the system stability index, achieving quantitative identification of abnormal correlations.

[0042] S4: Stability Assessment and Control Correction: Extract the verification results of the coupling relationship model from historical data of similar steel pile operations, and accumulate the deviation values ​​of each model over time to generate a system stability index. If the index continuously exceeds the preset trigger threshold, it is determined that control parameter correction needs to be performed, triggering the subsequent compensation mechanism.

[0043] The present invention will be further described below with reference to Examples 1 to 5:

[0044] Example 1:

[0045] In the data acquisition and cleaning stage, the hydraulic cylinder pressure data during steel pile operation is collected in real time using high-precision pressure sensors installed on the main and auxiliary hydraulic cylinders. The sampling frequency of the pressure sensors is set to 100Hz, with an accuracy class of 0.1, ensuring high accuracy of the collected pressure pulsation values ​​of the main hydraulic cylinder and the peak pressure values ​​of the auxiliary hydraulic cylinder. The steel pile displacement trajectory data is acquired by a multi-axis inertial measurement unit (IMU), which can simultaneously acquire triaxial acceleration and triaxial angular velocity data. The raw data is then processed using an extended Kalman filter algorithm to obtain the timing of the steel pile inclination change and the vertical displacement deviation. During the calculation process, by optimizing the filtering parameters and error correction model, the inclination calculation error is controlled within ±0.1°, and the vertical displacement deviation calculation error is controlled within ±2mm. The acquisition of geological environmental parameters is divided into two parts: soil shear strength parameters are obtained by sampling through on-site drilling and by standard testing methods such as direct shear test or triaxial compression test in the laboratory; the distribution of seawater impact force is calculated by combining the real-time data of wave height, period, and current velocity transmitted by the marine environmental monitoring station and using hydrodynamic formulas (such as Morrison's equation) to reflect the dynamic effect of seawater on steel piles under different sea conditions.

[0046] Data cleaning is a crucial step in ensuring the accuracy of subsequent analysis, encompassing four stages: noise filtering, data alignment, anomaly removal, and time-domain synchronization. For noise filtering, a second-order Butterworth low-pass filter is used to process the cylinder pressure and displacement trajectory data. The filter's cutoff frequency is set to 5Hz, a value determined based on the frequency range of the main signal components during steel pile operation (typically below 3Hz) and the noise frequency characteristics (primarily high-frequency interference). This filter effectively removes high-frequency noise mixed in during sensor acquisition, such as interference signals generated by equipment vibration, resulting in smoother pressure and displacement signals. The data alignment stage utilizes a synchronization trigger device to unify the timestamps of multi-source data. Specifically, a global clock source is set up in the steel pile operation system, synchronized with the pressure sensor, IMU, and geological data acquisition equipment, ensuring that each type of data carries a precise timestamp (millisecond-level accuracy) during acquisition. After data transmission to the processing system, a timestamp matching algorithm aligns data from different sources, ensuring that cylinder pressure, steel pile displacement, and geological parameters at the same moment correspond one-to-one, with a time synchronization error of less than 1ms.

[0047] Anomaly removal is achieved by setting reasonable thresholds and detection algorithms. For hydraulic cylinder pressure data, based on steel pile design parameters and historical operating experience, the normal range for main hydraulic cylinder pressure is set to 10-20 MPa, and the normal range for auxiliary hydraulic cylinder pressure is set to 5-15 MPa. Pressure values ​​exceeding these ranges are considered abnormal data points. Simultaneously, a displacement mutation detection algorithm is used to monitor the rate of change of the steel pile's inclination angle and vertical displacement. When the rate of change of the inclination angle exceeds 5° / s or the rate of change of the vertical displacement exceeds 10 mm / s, it is judged as a displacement mutation anomaly, and the corresponding data point is marked and removed. Furthermore, abnormal data in geological environmental parameters needs to be processed. For example, when the calculated results of seawater impact force distribution show values ​​that are significantly inconsistent with the current sea state characteristics (such as calculating abnormally large impact forces under windless weather), the data source and calculation process need to be checked, and invalid data needs to be removed. For time-domain synchronization processing, linear interpolation is used to fill in missing data points to address potential data loss or interruption issues during data acquisition. Specifically, when data loss is detected at a certain point in time, the missing value is calculated using a linear interpolation formula based on the values ​​and time intervals of adjacent data points before and after that point in time, thereby ensuring the continuity and integrity of the data sequence.

[0048] After data cleaning, the generated pressure and displacement feature sequences both contain timestamps, feature values, and feature labels. Each data point in the pressure feature sequence records the main cylinder pressure pulsation value, auxiliary cylinder pressure peak value, and feature label (e.g., normal, fluctuating) at the corresponding time. The displacement feature sequence includes the steel pile inclination angle, vertical displacement deviation, and feature label (e.g., stable, offset). Using timestamp indexing, the pressure and displacement feature sequences from different dimensions are dynamically matched to form a multi-source load feature set. This set, organized by time, integrates pressure, displacement, and geological parameters at the same moment into a structured data format, such as a table. Each row represents a time point, and each column corresponds to parameters such as timestamp, main cylinder pressure, auxiliary cylinder pressure, steel pile inclination angle, vertical displacement deviation, soil shear strength, and seawater impact force. This multi-source data integration method provides a unified input interface for subsequent feature extraction and model building, facilitating multi-dimensional analysis and correlation modeling.

[0049] Throughout the data acquisition and cleaning process, each step is closely interconnected and mutually influential. High-precision sensors and reliable data acquisition equipment are fundamental to obtaining raw data, ensuring that the data accurately reflects the actual state during steel pile operation. Appropriate noise filtering and data alignment methods effectively improve data quality and reduce the impact of noise and time deviations on the analysis results. Strict anomaly removal and time-domain synchronization processing guarantee data integrity and consistency, preventing deviations in model construction caused by abnormal data. Through this series of operations, the resulting multi-source load feature set lays a solid data foundation for subsequent steps such as dynamic load distribution parameter extraction and coupling relationship model construction. This enables the entire simulation analysis method to be modeled and verified based on reliable data, thereby accurately revealing the intrinsic relationship between the hydraulic cylinder force and the change in the steel pile's center of gravity.

[0050] Example 2:

[0051] In the coupling relationship model construction stage, feature parameters need to be extracted and screened from the cleaned multi-source data to establish a correlation model between the hydraulic cylinder force and the change of the steel pile's center of gravity. Taking the pile driving operation of a certain offshore wind power steel pile in a silty clay sea area as an example, hydraulic cylinder pressure, steel pile displacement, and geological parameter data are collected simultaneously during the operation. The pressure feature sequence includes the main hydraulic cylinder pressure pulsation value (such as periodic fluctuation data such as 12MPa, 13.5MPa, etc.) and the auxiliary hydraulic cylinder pressure peak value (such as reaching 18MPa during sudden impact). The displacement feature sequence records the change of the steel pile inclination angle (such as the initial vertical state inclination angle of 0°, which becomes 0.8° after 5 minutes of soil penetration) and the vertical displacement deviation (the cumulative deviation is about 50mm every 10 minutes). The geological environment parameters show that the soil shear strength is 30kPa, and the seawater impact force is affected by level 5 wind waves and fluctuates intermittently (peak value of about 20kN).

[0052] The extraction of load distribution characteristic parameters is based on sliding window analysis. Using a 5-minute window, the first derivative of the main cylinder pressure is calculated within the window to obtain the cylinder pressure gradient. For example, if the pressure increases from 12 MPa to 15 MPa within a 300-second window, the pressure gradient is 0.01 MPa / s. The second derivative is calculated by analyzing the time series of the steel pile inclination angle changes to obtain the center of gravity offset acceleration. Integration yields the center of gravity offset rate; for example, if the acceleration is 0.5 mm / s² within a certain time period. 2 The integrated velocity is 1.2 mm / s; the soil resistance coefficient is calculated based on the modified Terzaghi formula, and combined with the steel pile penetration depth of 10 meters and the soil shear strength of 30 kPa, the coefficient is 0.85.

[0053] The adaptive threshold setting combines historical data with real-time dynamic adjustment. First, 100 sets of operational data for similar steel piles in silty clay sea areas were retrieved from the historical database. The average hydraulic cylinder pressure gradient was 0.008 MPa / s with a standard deviation of 0.003 MPa / s, and the initial threshold was set at 0.008 ± 0.009 MPa / s (mean ± 3 times standard deviation). The average center of gravity offset rate was 1.0 mm / s with a standard deviation of 0.3 mm / s, and the threshold was set at 1.0 ± 0.9 mm / s. During real-time operation, a 10-minute sliding window was used, updating data every minute. At the 30-minute mark, soil resistance coefficients for three consecutive windows were detected as 1.2, 1.15, and 1.3 (exceeding the upper limit of the initial threshold of 0.85 ± 0.25, which is 1.1). The system automatically expanded the threshold to mean ± 4 times standard deviation (0.85 ± 0.32), retaining this parameter in the initial load feature set.

[0054] Cross-condition correlation analysis was conducted for different soil hardnesses during the piling stage. Three conditions were selected: soft soil (shear strength 15 kPa), plastic soil (30 kPa), and hard soil (60 kPa). Using plastic soil as the baseline, the cylinder pressure gradient deviation was calculated to be +0.005 MPa / s in the soft soil condition (baseline value 0.008 MPa / s), and -0.004 MPa / s in the hard soil condition. Pearson correlation coefficient analysis revealed a strong correlation between the cylinder pressure gradient and the center of gravity shift rate (0.82). The soil resistance coefficient was also found to be 0.65, close to the strong correlation threshold of 0.7, and therefore retained. The standard deviation of the deviations for each condition was calculated. The standard deviation of the pressure gradient in the soft soil condition was 0.002 MPa / s (25% of the mean), exceeding the preset stability threshold of 15%, and therefore included in the coupling relationship model.

[0055] The abrupt change characteristics of seawater impact force are supplemented through spectral analysis. When a 4Hz high-frequency component is detected in the seawater impact force spectrum (normal operating spectra are concentrated below 0.5Hz), it is determined to be a sudden wave impact. This abrupt change characteristic (e.g., the peak impact force suddenly increases from 20kN to 35kN) is extracted and input into the model. Combining pressure gradient, center of gravity shift rate, soil resistance coefficient, and impact force abrupt change parameters, a coupled relationship model is constructed. For example, a causal chain between parameters is established: sudden change in seawater impact force → fluctuation in soil resistance coefficient → adjustment of hydraulic cylinder pressure gradient → change in center of gravity shift rate.

[0056] After model construction, iterative verification is required. Taking hard soil conditions as an example, with input pressure gradient of 0.004 MPa / s, centroid shift rate of 0.8 mm / s, soil resistance coefficient of 1.5 (corrected), and impact force of 25 kN, the model predicts that the steel pile inclination angle change rate is 0.3° / min, which is within 5% of the actual monitored value of 0.28° / min, meeting the accuracy requirements. If the prediction error reaches 12% in a certain instance, the feature parameter selection process is checked back. It is found that the soil resistance coefficient in a certain window was incorrectly retained due to sensor malfunction (the actual value should be 1.2, but it was mistakenly sampled as 1.8). After removing this outlier, the model is retrained, and the error is reduced to 4%.

[0057] In another scenario, as the steel pile operation enters the stabilization phase, soil resistance fluctuations decrease (standard deviation is 10% of the mean), and cross-condition disturbance parameters are below the stability threshold. The model automatically eliminates the influence of inter-condition differences in soil resistance coefficients, retaining only the basic correlation between pressure gradient and center of gravity offset rate. At this point, if the seawater impact force tends to stabilize (no abrupt changes in the spectrum), the model focuses on the dynamic balance between cylinder pressure and center of gravity changes. For example, when the main cylinder pressure gradient is maintained at 0.002 MPa / s, the center of gravity offset rate should be stable within 0.5 mm / s. If the measured value reaches 0.9 mm / s, an abnormal correlation warning is triggered.

[0058] The entire coupling relationship model construction process extracts dynamic features through a sliding window, filters noise using adaptive thresholds, simplifies parameters based on correlation analysis, supplements environmental abrupt changes with physical meaning, and ensures model accuracy through iterative verification. Taking parameter changes in a specific operational scenario as an example, this process achieves a logical progression from raw data to structured features, and from single-parameter analysis to multi-parameter coupling. This enables the model to accurately reflect the intrinsic mechanisms of load distribution and center of gravity changes during steel pile operations, providing a reliable mathematical foundation for subsequent dynamic verification and control correction.

[0059] Example 3:

[0060] During the dynamic verification and anomaly identification phase, the system uses a real-time monitoring module to acquire the rate of change of hydraulic cylinder pressure and the acceleration of the steel pile's center of gravity offset, enabling precise analysis of the dynamic correlation between hydraulic cylinder pressure and steel pile displacement. The real-time monitoring module is designed based on a hardware-triggered interrupt mechanism. When an update to hydraulic cylinder pressure or steel pile displacement data is detected, an interrupt signal is immediately triggered, and the current rate of change of hydraulic cylinder pressure is simultaneously acquired. and the acceleration of the center of gravity shift of the steel pile The data acquisition period is set to 100ms to ensure data real-time performance and synchronization. Among other things, the hydraulic cylinder pressure change rate... It is obtained by numerical differentiation of the pressure characteristic sequence, specifically using the forward difference method, as shown in the formula:

[0061]

[0062] In the formula, For the current moment The hydraulic cylinder pressure value (unit: MPa). For the previous moment The hydraulic cylinder pressure value (unit: MPa). The sampling interval (unit: s) is 0.1s (i.e., 100ms). Acceleration due to center of gravity shift of the steel pile. This is obtained by differentiating the centroid offset rate in the displacement characteristic sequence, or directly by obtaining triaxial acceleration data calculated by a multi-axis inertial measurement unit (IMU) through coordinate transformation and vector synthesis, with units of [unit missing]. .

[0063] Calculating the dynamic response difference is a core step in determining abnormal mechanical correlations. Define the dynamic response difference. The rate of change of cylinder pressure acceleration due to shift of the center of gravity of the steel pile The difference in absolute values, that is:

[0064]

[0065] The preset safety range is determined by analyzing the statistical characteristics of historical operation data. Specifically, 1000 sets of normal operating conditions are selected from the historical database. and Data, calculate each set of data Values, construct their probability distribution model (such as normal distribution), and use the mean Add or subtract 1.96 times the standard deviation This serves as the upper and lower limits of a preset safety range (covering 95% of normal data). When calculated in real-time... If the value exceeds this range, it is determined to be an abnormal mechanical correlation, indicating that the coupling relationship between the force on the hydraulic cylinder and the change in the center of gravity of the steel pile is abnormal and requires further analysis.

[0066] The statistical analysis of abnormal mechanical correlations and the acquisition of soil resistance fluctuation amplitudes are carried out simultaneously. This is done within a preset operational cycle. Within a 30-minute period (e.g., the system automatically counts the number of abnormal mechanical correlations that occur, and records them as follows: Fluctuation range of soil resistance This was obtained by analyzing the soil resistance coefficient sequence, specifically by calculating the maximum value of the soil resistance coefficient within the work cycle. and minimum value The difference is:

[0067]

[0068] The soil resistance coefficient is calculated by combining the modified Terzaghi formula with real-time soil shear strength parameters and the depth of steel pile penetration, and the unit is a dimensionless quantity or a specific engineering unit (such as kPa·m).

[0069] The system stability index is generated through a dynamic correlation function, which is based on the number of abnormal mechanical correlations. As input variables, the amplitude of soil resistance fluctuations To correct for variables, a prediction model based on multibody dynamics is used to achieve parameter mapping. The prediction model is constructed based on the finite element simulation algorithm, and a multiphysics coupled simulation model of steel pile-hydraulic cylinder-soil is established to simulate different... and The system response under combined conditions. The model input layer contains standardized... and The value, standardized by the formula is:

[0070]

[0071] In the formula, and These represent the mean and standard deviation of the number of anomalous mechanical correlations in historical data, respectively. and These represent the mean and standard deviation of soil resistance fluctuations, respectively. The output layer represents the system stability index. The value ranges from 0 to 100, with larger values ​​indicating poorer system stability. The hidden layer of the model uses a multilayer perceptron structure, and the number of neurons is determined based on the complexity of the simulation data (e.g., 10-20). The ReLU activation function is used to improve computational efficiency.

[0072] The training process of the dynamic correlation function is as follows: First, 1000 different sets of functions are generated through finite element simulation. and Combined sample data, each sample group contains input parameters. and corresponding system stability indicators Then, the sample data is divided into a training set (800 sets) and a test set (200 sets). The backpropagation algorithm is used to optimize the model parameters (such as weights and biases) so that the prediction error (root mean square error) of the training set is less than a preset threshold (such as 5). Finally, the generalization ability of the model is verified through the test set to ensure that the model can still accurately output stability indicators on unseen data.

[0073] When system stability indicators When the failure threshold is exceeded (e.g., 80), it indicates that the system faces a high risk of instability and requires triggering a control parameter correction command. The determination of the preset failure threshold combines safety regulations for steel pile operations with historical failure data. For example, the critical stability index value that led to system instability in historical data is selected as the threshold to ensure that the compensation mechanism is activated in time before the risk occurs.

[0074] The entire dynamic verification and anomaly identification process, through real-time data acquisition, difference calculation, statistical analysis, and model prediction, achieves quantitative identification of the correlation between hydraulic cylinder pressure fluctuations and abnormal steel pile center of gravity offset. A hardware-triggered interrupt mechanism ensures the synchronization and real-time nature of data acquisition, while numerical differential algorithms and statistical methods ensure the accuracy of parameter calculations. The predictive model trained by finite element simulation provides a reliable mathematical basis for stability assessment. Each stage is tightly connected through data flow, forming a closed-loop control system from data acquisition to anomaly detection, providing key technical support for the dynamic monitoring and stability management of the steel pile operation process.

[0075] Example 4:

[0076] When the system stability index exceeds the preset trigger threshold (e.g., 70) for three consecutive operating cycles (each cycle is set to 30 minutes), the system determines that control parameter correction is required and automatically initiates the cylinder pressure compensation scheme generation process. The core objective of the compensation scheme is to restore the coupled stability between the cylinder force and the change in the center of gravity of the steel pile by adjusting the cylinder pressure parameters and the center of gravity offset control parameters. The scheme includes three key parameters: the main cylinder pressure adjustment gradient, the auxiliary cylinder pressure limit value, and the center of gravity offset compensation coefficient. The generation and optimization of each parameter are based on a multi-constraint optimization algorithm to ensure a balance between energy consumption, stability, and response speed under different operating scenarios.

[0077] Taking the piling operation of a certain type of steel pile in a plastic soil seabed environment as an example, when the system stability index remains at 75 for 90 consecutive minutes (exceeding the trigger threshold of 70), the compensation mechanism is triggered. First, a multi-constraint optimization function is used to construct an optimization problem with the objectives of minimizing cylinder energy consumption, maximizing center of gravity offset suppression, and optimizing system response speed. Minimizing cylinder energy consumption aims to reduce energy consumption during operation, which is achieved by controlling the pressure output of the main and auxiliary cylinders; maximizing center of gravity offset suppression requires quickly correcting the center of gravity offset of the steel pile by adjusting the pressure distribution; optimizing system response speed emphasizes the timeliness of pressure adjustment to avoid instability exacerbation due to delay.

[0078] During parameter optimization, a particle swarm optimization (PSO) algorithm is used to iteratively calculate three key parameters. At algorithm initialization, a population of 50 particles is set, with each particle representing a set of parameters: the main cylinder pressure adjustment gradient (range ±5 MPa / min), the auxiliary cylinder pressure limit (range 10-18 MPa), and the center of gravity offset compensation coefficient (range 0.1-0.5). The inertia weight is linearly decreased from 0.9 to 0.4, and the learning factor is set to 2 to balance global search and local exploitation capabilities. During iteration, each particle updates its trajectory based on its current fitness (calculated by a weighted sum of the objective functions) and its historical best position, gradually approaching the optimal solution.

[0079] The dynamic adjustment of weighting coefficients is based on the needs of the current operational stage. For example, in the initial stage of pile driving, the goal is rapid soil penetration, with a weighting of 60% for center of gravity offset suppression, 30% for energy consumption, and 10% for response speed. When the steel pile reaches 80% of the design depth, the stage transitions to pile stabilization, increasing the energy consumption weight to 50%, adjusting the center of gravity offset suppression weight to 40%, and maintaining the response speed weight at 10%. This dynamic adjustment mechanism allows the algorithm to adapt to the control objectives at different stages, avoiding system performance imbalances caused by optimizing a single objective.

[0080] After 100 iterations, the algorithm outputs a non-dominated solution set (Pareto solution set), containing 3 sets of candidate compensation schemes:

[0081] Option 1: Main cylinder pressure adjustment gradient +3MPa / min, auxiliary cylinder pressure limit 15MPa, center of gravity offset compensation coefficient 0.3;

[0082] Option 2: Main cylinder pressure adjustment gradient +2MPa / min, auxiliary cylinder pressure limit 16MPa, center of gravity offset compensation coefficient 0.4;

[0083] Option 3: Main cylinder pressure adjustment gradient +4MPa / min, auxiliary cylinder pressure limit 14MPa, center of gravity offset compensation coefficient 0.2.

[0084] To verify the robustness of the candidate schemes, Monte Carlo simulation was used to generate 1000 sets of different sea state parameter combinations, covering wave heights of 0.5-3m, current velocities of 0.2-1.5m / s, and soil resistance fluctuations of ±20%. Taking Scheme 1 as an example, the simulation was conducted under extreme conditions of wave height of 2m, current velocity of 1m / s, and a sudden increase in soil resistance of 15%. Its stability index was calculated to be 68 (below the safety threshold of 70), indicating that the scheme can still maintain system stability under highly volatile environments. By statistically analyzing the mean and variance of the stability index of each scheme under all simulated conditions, Scheme 2, with the lowest mean and smallest variance (mean 65, variance 8.2), was selected as the highest priority execution scheme. Schemes 1 and 3 were selected as secondary and tertiary alternative schemes, respectively.

[0085] The execution priority of the compensation scheme is determined according to the degree of deviation of the system stability index. If the index is in the 70-80 range (Level 1 alarm), Scheme 1, which balances energy consumption and response speed, is executed first. If the index exceeds 80 (Level 2 alarm), Scheme 2, which has the strongest center of gravity offset suppression, is activated. In the example above, since the index is 75 (Level 1 alarm), the system first executes Scheme 1: the pressure of the main hydraulic cylinder is gradually increased in a gradient of 3 MPa / min, while the upper limit of the auxiliary hydraulic cylinder pressure is limited to 15 MPa. The pressure distribution of each hydraulic cylinder is adjusted by a center of gravity offset compensation coefficient of 0.3 to suppress the center of gravity offset of the steel pile to the right (offset rate of 2.5 mm / s).

[0086] During execution, the real-time monitoring module continuously collects data on cylinder pressure and steel pile displacement, calculating stability indicators every 5 minutes. If, after implementing Scheme 1, the indicator drops to 68 (below the threshold of 70) within the first cycle (30 minutes), the current parameters are maintained; if the indicator does not drop significantly (e.g., remains at 73), it automatically switches to Scheme 2, increasing the auxiliary cylinder pressure limit to 16MPa and raising the center of gravity offset compensation coefficient to 0.4, thereby enhancing the control over center of gravity offset.

[0087] The entire compensation scheme generation and execution process closely integrates real-time data and historical experience. Multi-objective optimization algorithms ensure the effectiveness of parameter combinations, Monte Carlo simulations verify the scheme's environmental adaptability, and a graded response mechanism enhances the system's emergency response capabilities. This data-driven and algorithm-optimized control strategy can rapidly generate targeted compensation schemes when steel pile operations face complex loads and environmental disturbances, achieving a dynamic balance between the hydraulic cylinder force and the steel pile's center of gravity changes, thus ensuring the safety and stability of the operation process.

[0088] Example 5:

[0089] The conflict balancing mechanism of the multi-constraint optimization function is achieved through dynamic weight adjustment. This mechanism is based on a fuzzy logic controller design. The input variables are the system stability index, the hydraulic cylinder pressure adjustment margin, and the center of gravity offset rate. The output is the weight coefficient of each objective function. Taking the pile stabilization operation of a certain type of steel pile in a hard soil seabed environment as an example, when the system stability index is 85 (higher than the preset failure threshold of 80), the hydraulic cylinder pressure adjustment margin is 15% (the current pressure is close to the maximum allowable pressure), and the center of gravity offset rate is 4 mm / s (exceeding the normal range of 2 mm / s), the fuzzy logic controller adjusts the weights according to the following rules:

[0090] If the stability index is high and the center of gravity shift rate is high, then increase the weight of the center of gravity shift suppression.

[0091] If the hydraulic cylinder pressure adjustment margin is low, reduce the hydraulic cylinder pressure adjustment gradient to avoid overload.

[0092] Based on the input variables, the objective function is automatically generated with the following weights: energy consumption 20%, center of gravity shift suppression 70%, and response speed 10%.

[0093] During the Monte Carlo simulation, random samples covering different sea states were generated for candidate pressure compensation schemes. For example, a scenario with a wave height of 1.5m, a current velocity of 0.8m / s, and soil resistance fluctuation of +10% was simulated to verify the stability of the scheme under this scenario. Taking a set of parameter combinations (main cylinder pressure adjustment gradient +2MPa / min, auxiliary cylinder pressure limit value of 16MPa, and center of gravity offset compensation coefficient of 0.4) as an example, the stability index distribution in 1000 random samples was simulated and calculated:

[0094] Wave height 0.5-1m range: indicator average 62, fluctuation range 55-68;

[0095] Wave height 1-2m range: indicator average 70, fluctuation range 63-78;

[0096] For waves 2-3m high: the average index is 82, with a fluctuation range of 75-89 (exceeding the failure threshold of 80 20 times). Statistical analysis shows that the stability index of this scheme is below 80 in 85% of the samples, and its robustness meets the requirements (the preset robustness probability threshold is 85%), therefore it is retained as a valid candidate scheme.

[0097] The specific implementation of the dynamic weight adjustment mechanism includes the definition of fuzzy rules and the design of membership functions. The membership function divides the input variables into multiple fuzzy sets, for example:

[0098] System stability index: {Low (<60), Medium (60-75), High (>75)};

[0099] Hydraulic cylinder pressure adjustment margin: {Sufficient (>30%), Medium (15%-30%), Insufficient (<15%)};

[0100] Center of gravity offset rate: {slow (<2mm / s), medium (2-4mm / s), fast (>4mm / s)}. For different combinations of fuzzy sets, corresponding weight adjustment rules are predefined. For example, when the stability index is "high," the pressure regulation margin is "insufficient," and the center of gravity offset rate is "fast," the rule specifies that the center of gravity offset suppression weight is increased to 70%, the energy consumption weight is reduced to 20%, and the response speed weight is maintained at 10%, in order to prioritize controlling the center of gravity offset while avoiding cylinder overload.

[0101] In another example of pile stabilization operation, the system stability index was 72 (Level 1 alarm), the cylinder pressure adjustment margin was 25% (moderate), and the center of gravity offset rate was 1.8 mm / s (slow). At this point, the fuzzy logic controller determined that the primary risk was excessive energy consumption, and therefore adjusted the weights to: energy consumption 50%, center of gravity offset suppression 30%, and response speed 20%. The particle swarm optimization algorithm re-optimized based on the new weights, generating the following parameter combination: main cylinder pressure adjustment gradient +1 MPa / min, auxiliary cylinder pressure limit 14 MPa, and center of gravity offset compensation coefficient 0.2. In Monte Carlo simulations, under conditions of 1 m wave height and 0.5 m / s flow velocity, this scheme achieved an average stability index of 65, and the cylinder pressure fluctuation amplitude decreased by 12%, demonstrating the guiding role of weight adjustment in energy consumption optimization.

[0102] The robustness-validated Latin hypercube sampling method ensures that the sample covers the entire parameter space. Taking soil resistance fluctuations as an example, the sampling points are evenly distributed within the range of -20% to +20%, avoiding the clustering bias of traditional random sampling. For each candidate scheme, its stability index is calculated across all sampling points, and the probability of it falling below a safety threshold (e.g., 70) is statistically analyzed. If a scheme has an index of 72 (exceeding the threshold) when soil resistance fluctuates by +15%, and the probability of this fluctuation is 10%, then the robustness probability of the scheme is 90% (meeting the threshold of 85%), and it can be retained; if the probability is only 80%, the algorithm is returned to readjust the parameters.

[0103] The synergistic effect of dynamic weights and robustness verification is reflected in the fact that when weight adjustments cause candidate solutions to be biased towards a certain objective (such as center of gravity shift suppression), robustness simulation can test the performance of the solution in other objective dimensions. For example, a solution may have a significant effect on center of gravity shift suppression under high weights, but Monte Carlo simulations show that its performance fluctuates greatly in energy-sensitive scenarios (such as long-term operation). In this case, by fine-tuning the weights (such as increasing the energy consumption weight by 10%) and re-optimizing, a new solution that takes into account multiple objectives can be generated.

[0104] Under complex sea conditions (e.g., wave height 2.5m, current velocity 1.2m / s, soil resistance fluctuation -10%), the initial stability index of a candidate scheme was 88 (level 2 alarm). After dynamic weight adjustment, the weight for suppressing center of gravity offset increased from 60% to 80%. The algorithm generated new parameters: main cylinder pressure adjustment gradient +4MPa / min, auxiliary cylinder pressure limit 18MPa (close to the maximum allowable value), and center of gravity offset compensation coefficient 0.5. Monte Carlo simulation showed that the average index of this scheme decreased to 78 under similar extreme conditions. Although still higher than the failure threshold, it was 10 units lower than the original scheme, indicating that the weight adjustment effectively improved the scheme's relevance.

[0105] Throughout the implementation process, the dynamic weight adjustment mechanism achieves intelligent balancing of multi-objective conflicts through fuzzy logic, while Monte Carlo simulation and Latin hypercube sampling ensure the reliability of the solution across the entire operating range. The combination of these two methods forms a closed-loop process of "optimization-verification-re-optimization." This mechanism enables the stress compensation scheme to dynamically adjust its optimization direction based on real-time operating conditions and environmental parameters. This avoids the limitations of single-objective optimization and ensures the robustness of the scheme in complex environments, providing a flexible and reliable solution for the stability control of steel pile operation systems.

[0106] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0107] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A steel pile oil cylinder stress and steel pile gravity center change model simulation analysis method, characterized in that, The method comprises the following steps: S1: acquiring cylinder pressure data, steel pile displacement trajectory data and geological environment parameters in the steel pile operation process, and performing data cleaning on the cylinder pressure data, displacement trajectory data and geological parameters; S2: performing multi-dimensional feature extraction on the cleaned data to generate dynamic load distribution parameters, and constructing a coupling relationship model of cylinder stress and steel pile gravity center change based on the load distribution parameters; S3: according to the coupling relationship model, the real-time dynamics of the cylinder pressure and the steel pile displacement are jointly verified, the spatial correspondence of the cylinder pressure gradient change and the steel pile gravity center offset is identified through multi-axis motion correlation technology, and the abnormal association of the cylinder pressure fluctuation and the gravity center offset is quantitatively identified; S4: generating a system stability index according to the verification result of the coupling relationship model, and judging whether to trigger control parameter correction based on the system stability index.

2. The method according to claim 1, wherein the method is characterized in that: The cylinder pressure data includes main cylinder pressure pulsation value and auxiliary cylinder pressure peak value; the displacement trajectory data includes steel pile inclination angle change time sequence and vertical displacement deviation; the geological environment parameters include soil shear strength parameters and seawater impact force distribution; the data cleaning of the cylinder pressure data, displacement trajectory data and geological parameters includes noise filtering, data alignment, abnormal event elimination and time domain synchronization processing; the cylinder pressure data cleaning generates a pressure feature sequence, and the displacement trajectory data cleaning generates a displacement feature sequence; the pressure feature sequence and the displacement feature sequence of different dimensions are dynamically matched to form a multi-source load feature set.

3. The method according to claim 2, wherein the method is characterized in that: In S2, the following steps are included: S201: extracting load distribution feature parameters from the cleaned pressure feature sequence, displacement feature sequence and geological environment parameters, respectively, the load distribution feature parameters including cylinder pressure gradient, gravity center offset rate and soil resistance coefficient, combining each load distribution feature parameter according to a preset correlation degree to generate dynamic load distribution parameters; S202: setting an adaptive threshold for the dynamic load distribution parameters in the multi-source load feature set, screening feature parameters meeting the threshold range to form an initial load feature set, and excluding feature parameters exceeding the threshold range; S203: performing cross-condition correlation analysis on each feature parameter in the initial load feature set, extracting the deviation of the feature parameters in different conditions in the same operation stage, calculating the standard deviation of the deviation and marking it as a inter-condition disturbance parameter; S204: comparing the inter-condition disturbance parameter with a preset stability threshold, screening feature parameters exceeding the stability threshold to add to the coupling relationship model, and supplementing seawater impact force mutation features to the coupling relationship model according to the geological environment parameters.

4. The method according to claim 3, characterized in that: In S3, the abnormal association of the cylinder pressure fluctuation and the gravity center offset is quantitatively identified, comprising the following steps: S301: real-time monitoring of the cylinder pressure change rate and the steel pile gravity center offset acceleration, calculating the dynamic response difference value, and determining abnormal mechanical association if the dynamic response difference value exceeds a preset safety interval; S302: counting the number of abnormal mechanical associations in a preset operation period as N, and synchronously acquiring the soil resistance fluctuation amplitude S in the geological environment parameters; S303: According to the interaction relationship between the abnormal mechanical correlation number N and the soil resistance fluctuation amplitude S, a dynamic correlation function is used to generate a system stability index, wherein the product component and the difference component of N and S are generated by composite operation to generate a comprehensive evaluation value; if the system stability index exceeds the preset failure threshold, a control parameter correction instruction is triggered.

5. The method according to claim 4, wherein the method is characterized in that: In S4, the verification results of the coupling relationship model in the same type of steel pile operation historical data are extracted, the time series accumulation of the deviation value of each model is generated to generate a system stability index, and if the system stability index continuously exceeds the preset trigger threshold, it is determined that control parameter correction needs to be performed.

6. The method according to claim 3, characterized in that: In S202, the setting method of the adaptive threshold includes: based on the dynamic distribution range of each characteristic parameter in the historical operation database, the fluctuation interval of the characteristic parameter is calculated by using the sliding window algorithm, the upper limit and the lower limit of the fluctuation interval are taken as the reference value of the adaptive threshold, and the threshold range is dynamically corrected according to the characteristic parameter distribution of real-time operation data.

7. The method according to claim 4, characterized in that: In S303, the specific implementation of the dynamic correlation function is: a composite function is established with the abnormal mechanical correlation number N as the input variable and the soil resistance fluctuation amplitude S as the correction variable, a prediction model of the system stability index is obtained by training through a finite element simulation algorithm, the prediction model adopts a multi-body dynamics structure, the input layer includes the standardized values of N and S, and the output layer is the system stability index.

8. The method according to claim 1, wherein the method is characterized in that: Further comprising S5: when it is determined that control parameter correction needs to be triggered, an oil cylinder pressure compensation scheme is automatically generated, the pressure compensation scheme includes a main oil cylinder pressure adjustment gradient, an auxiliary oil cylinder pressure limiting value and a gravity center offset compensation coefficient, and the execution priority of the compensation scheme is graded according to the deviation degree of the system stability index.

9. The method according to claim 8, characterized in that: In S5, the generation method of the pressure compensation scheme includes: constructing a multi-constraint optimization function, taking the minimization of oil cylinder energy consumption, the maximization of gravity center offset inhibition and the optimization of system response speed as constraint conditions, using a particle swarm optimization algorithm for multi-parameter optimization calculation, and outputting a non-dominated solution set as a candidate pressure compensation scheme.

10. The method according to claim 9, wherein the method is characterized in that: The multi-constraint optimization function verifies the robustness of the solution set by Monte Carlo simulation, balances the conflict relationship between each constraint condition by using a dynamic weight adjustment mechanism, and ensures that the candidate pressure compensation scheme meets the operation requirements under different sea conditions.

Images