Method and system for predicting stress response and displacement response of jacket platform

Abstract

The invention discloses a method and a system for predicting stress response and displacement response of a jacket platform, and belongs to the technical field of health monitoring, and the method comprises the following steps: establishing a finite element model of the jacket platform, and obtaining the inherent frequency and local strain of the jacket platform through simulation; correcting the finite element model of the jacket platform by adjusting model parameters of the finite element model by respectively taking errors corresponding to the simulated and real-time acquired inherent frequency and local strain as variables; taking real wind wave load data as excitation, inputting the real wind wave load data into the finite element correction model of the jacket platform, obtaining stress and displacement response data of the jacket platform through simulation, and respectively constructing a stress response reduced-order model and a displacement response reduced-order model according to optimal modal basis vectors and modal coefficients corresponding to different data sets, and stress response and displacement response of the in-service jacket platform are monitored in real time. According to the method, finite element model correction and model order reduction technologies are combined, so that high-precision prediction of jacket platform stress and displacement response is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of health detection, more particularly to a method and system for predicting stress response and displacement response of a jacket platform. BACKGROUND

[0002] The structural integrity of a jacket platform of marine oil and gas equipment is related to national energy security and marine ecological protection. However, a large number of in-service and over-service jacket platforms still do not have a structural health monitoring system, making it difficult to realize real-time sensing and early warning of structural responses under complex marine loads such as wind, wave and current.

[0003] Based on meteorological and sea state data in the area where the jacket platform structure is located, combined with intelligent algorithm models, the dynamic response of the jacket platform structure can be predicted online under the condition of lacking a traditional structural health monitoring system, and a safety warning can be issued in time when potential abnormal conditions are identified, which is of great significance for reducing the risk of accidents, prolonging the service life of the structure, and protecting personnel and environmental safety.

[0004] At present, the Chinese invention patent with the patent number CN118395808A discloses a jacket platform digital twin method and device, which obtains the load of the monitoring point by using the load inversion method and applies the inverted load on the finite element jacket to obtain the stress and displacement information of the structure.

[0005] However, for complex working conditions, due to the uncertainty of the jacket platform site environment and working conditions, the static data set obtained by the prior art cannot accurately construct a high-fidelity finite element model, and the error between the predicted results and the actual results is too large, which cannot reflect the structural stress and displacement under some working conditions. SUMMARY

[0006] In view of the problems existing in the above field, the present application proposes a method and system for predicting stress response and displacement response of a jacket platform, which corrects the finite element model through multi-source data fusion, and combines the finite element correction model with modal reduction technology, constructs stress response reduction model and displacement response reduction model respectively, and realizes high-precision prediction of the stress and displacement response of the jacket platform.

[0007] To solve the above technical problems, the present application discloses a method for predicting stress response and displacement response of a jacket platform structure, comprising the following steps: Obtain the real structure parameters, real wind and wave load data, acceleration and strain monitoring data of the in-service jacket platform to be measured; Based on the real structure parameters, a jacket platform finite element model is established, and the natural frequency and local strain of the jacket platform are obtained by simulation; based on the acceleration and strain monitoring data, the natural frequency and local strain of the jacket platform are obtained in real time; taking the error of the natural frequency and local strain obtained by simulation and in real time as variables, the target value of the finite element model correction is determined by constructing a target function; by adjusting the model parameters of the finite element model, when the error of the target value meets the error threshold requirement, the jacket platform finite element correction model is obtained; The real wind and wave load data are input into the jacket platform finite element correction model as excitation, and the stress response data set and displacement response data set of the jacket platform are obtained by simulation; The optimal modal basis vector corresponding to the stress response data set and displacement response data set and the modal coefficients corresponding thereto are obtained; the modal coefficients corresponding to the optimal modal basis vector corresponding to different data sets are predicted, and the optimal modal basis vector corresponding to different data sets and the predicted modal coefficients are linearly combined and superimposed to construct a stress response reduction model and a displacement response reduction model, respectively, to monitor the stress response and displacement response of the in-service jacket platform in real time.

[0008] Preferably, the optimal modal basis vector corresponding to the stress response data set and displacement response data set and the modal coefficients corresponding thereto are obtained, specifically including: Based on the intrinsic orthogonal decomposition method, the modal basis vector corresponding to the stress response data set Q and the modal basis vector corresponding to the displacement response data set W and the modal coefficients corresponding thereto are determined; According to the modal basis vector corresponding to the stress response data set Q and the modal basis vector corresponding to the displacement response data set W and the modal coefficients corresponding thereto, the optimal modal basis vector corresponding to the stress response data set Q and the displacement response data set W and the modal coefficients corresponding thereto are determined based on the leave-one-out cross-validation method.

[0009] Preferably, the optimal modal basis vector corresponding to the stress response data set and displacement response data set and the modal coefficients corresponding thereto are obtained, specifically including: According to the optimal modal basis vector corresponding to the stress response data set Q and the modal coefficients corresponding thereto, the real wind and wave load data and the previous nThe modal coefficients corresponding to the optimal modal basis vectors of the stress response data set are input into the LSTM neural network as training data, the hyperparameters of the LSTM neural network, including the number of hidden layer nodes, the learning rate and the time step, are optimized using the grey wolf optimization algorithm, and the modal coefficients corresponding to the optimal modal basis vectors of the stress response data set are predicted Q The optimal modal basis vectors of the stress response data set Q are linearly combined with the predicted modal coefficients and superimposed to construct a stress response reduced-order model for real-time monitoring of the stress response of the in-service jacket platform. According to the optimal modal basis vectors of the displacement response data set W and the modal coefficients corresponding thereto, the real wave load data and the previous m The modal coefficients corresponding to the optimal modal basis vectors of the displacement response data set are input into the radial basis neural network as training data, the hyperparameters of the radial basis neural network, including the number of hidden layer nodes, the basis function width and the regularization coefficient, are optimized using the sparrow search algorithm, and the modal coefficients corresponding to the optimal modal basis vectors of the displacement response data set are predicted W The optimal modal basis vectors of the displacement response data set W are linearly combined with the predicted modal coefficients and superimposed to construct a displacement response reduced-order model for real-time monitoring of the displacement response of the in-service jacket platform.

[0010] Preferably, the real wave load data is input into the jacket platform finite element correction model as an excitation, and the stress response data set and the displacement response data set of the jacket platform are obtained by simulation, specifically including: The real wave load data includes time history data of wind load and wave load; Based on the wind speed sensor and the wave sensor installed on the in-service jacket platform, real-time wind speed and wind direction information is obtained; According to the wind speed and wind direction information, a wind pressure time history is constructed to obtain the time history data of the wind load, which is applied to the upper structure surface of the jacket platform finite element correction model; Based on the Morison equation, the time history data of the wave load is generated, and the time history data of the wave load is applied to the underwater part of the jacket platform finite element correction model in the form of nodes or elements; In the ANSYS environment, transient dynamics simulation analysis is performed on the jacket platform finite element correction model, the Newmark time integration method is used as the core algorithm of the solver of the jacket platform finite element correction model, the time step is set to 0.1s, and the platform pipe nodes are taken as the research object to obtain the stress and displacement response data of the jacket platform; According to the obtained stress response data and displacement response data of the jacket platform, the stress response data set Q and the displacement response data set W are constructed.

[0011] Preferably, the finite element model of the jacket platform is established based on the real structure parameters, and the natural frequency and the local strain of the jacket platform are obtained by simulation, and specifically comprising: The real structure parameters of the in-service jacket platform include material properties, size parameters and geometric characteristics; Based on the material properties, size parameters and geometric characteristics of the in-service jacket platform, a finite element model of the jacket platform is established by using ANSYS software, wherein the pile leg and the chord diagonal brace structure of the finite element model are established by using beam188 element, the deck of the finite element model is established by using shell181 element, and the pipe node of the splash zone of the deck is established by using soild187 element; The natural frequency and the local strain of the jacket platform are obtained by numerical simulation based on the finite element technology.

[0012] Preferably, the error corresponding to the natural frequency and the local strain obtained by simulation and in real time is taken as a variable, the target value of the finite element model correction is determined by constructing a target function, and specifically comprising: The constructed target function is: ; In the formula: E is the output of the target function, that is, the target value of the finite element model correction; is the natural frequency of the jacket platform obtained by numerical simulation; is the natural frequency of the in-service jacket platform obtained by real-time monitoring; is a weight coefficient; is the local strain of the finite element model of the jacket platform obtained by numerical simulation; is the local strain of the in-service jacket platform obtained by real-time monitoring.

[0013] Preferably, the jacket platform finite element correction model is obtained by adjusting the model parameters of the finite element model when the error of the target value meets the error threshold requirement, and specifically comprising: The model correction strategy of the finite element model is obtained by combining the Kriging surrogate model with the multi-objective genetic algorithm, the model parameters of the finite element model including the material density, the elastic modulus and the wall thickness of the pipe wall are adjusted, the target value to be corrected is iteratively optimized, and the jacket platform finite element correction model is obtained when the error between the target value of the jacket platform finite element model and the natural frequency and the local strain of the in-service jacket platform obtained by real-time monitoring meets the error threshold requirement.

[0014] Preferably, the natural frequency and the local strain of the jacket platform are obtained in real time based on the acceleration and strain monitoring data, and specifically comprising: Based on the vibration monitoring experiment, the acceleration and strain monitoring data are collected by the sensing device; The acceleration and strain data are preprocessed, including detrending, filtering and noise removal; Based on the preprocessed acceleration and strain data, the natural frequency and local strain of the in-service jacket platform are obtained.

[0015] Preferably, the stress response and displacement response prediction system of the jacket platform further comprises: The data acquisition module is used to acquire the real structure parameters, real wind and wave load data, acceleration and strain monitoring data of the in-service jacket platform to be measured; The finite element model correction module is used to establish the finite element model of the jacket platform based on the real structure parameters, to obtain the natural frequency and local strain of the jacket platform by simulation, to obtain the natural frequency and local strain of the jacket platform in real time based on the acceleration and strain monitoring data, to determine the target value of the finite element model correction by constructing the objective function with the errors of the natural frequency and local strain obtained by simulation and in real time as variables, and to obtain the finite element correction model of the jacket platform by adjusting the model parameters of the finite element model when the error of the target value meets the error threshold requirement; The finite element simulation module is used to input the real wind and wave load data as excitation into the finite element correction model of the jacket platform, and to obtain the stress response data set and displacement response data set of the jacket platform by simulation; The prediction module is used to obtain the optimal modal basis vector and the modal coefficient corresponding thereto corresponding to the stress response data set and displacement response data set, to predict the modal coefficient corresponding to the optimal modal basis vector corresponding to different data sets, to linearly combine and superimpose the optimal modal basis vector corresponding to different data sets and the predicted modal coefficient, and to construct the stress response reduced-order model and displacement response reduced-order model respectively, so as to monitor the stress response and displacement response of the in-service jacket platform in real time.

[0016] Compared with the prior art, the present application has the following beneficial effects: This invention proposes a method for predicting the stress and displacement responses of a jacket platform. This method comprehensively utilizes multi-source information, including the actual structural parameters, acceleration, and strain monitoring data of the in-service jacket platform under test. It can more comprehensively and accurately reflect the actual state of the jacket platform. Compared with correction methods based on a single data source, it can more effectively reduce the difference between the finite element model and the actual structure of the jacket platform, significantly improving the accuracy of the finite element model and laying a solid foundation for subsequent accurate stress and displacement response predictions. By setting an error threshold, the corrected model is ensured to have high accuracy, avoiding prediction errors caused by insufficient or excessive model correction, further improving the reliability and usability of the model. Using real wind and wave load data as excitation input into the corrected finite element model can more realistically simulate the stress conditions of the jacket platform under actual working conditions, making the obtained response data closer to the actual situation, thereby improving the accuracy of subsequent predictions. Modal analysis-based methods can capture the main characteristics and dynamic properties of the jacket platform response. By predicting modal coefficients, the adaptability to response changes can be further improved. By linearly combining and superimposing the optimal modal basis vectors and predicted modal coefficients to construct a reduced-order model, the stress response and displacement response can be described more accurately while reducing computational complexity, thus enhancing the accuracy of the prediction results. Attached Figure Description

[0017] Figure 1 This is a flowchart of the stress response and displacement response prediction method for the jacket platform proposed in this invention; Figure 2 This invention provides a stress response and displacement response prediction system for a jacket platform. Detailed Implementation

[0018] The following will refer to the appendices in the embodiments of the present invention. Figures 1-2 The technical solutions in the embodiments of the present invention will be clearly and completely described. It should be understood that the terminology used in the present invention is only for describing particular implementation methods and is not intended to limit the present invention.

[0019] Example like Figure 1 As shown, this invention proposes a method for predicting the stress response and displacement response of a jacket platform, comprising the following steps: S1: Obtain the actual structural parameters, actual wind and wave load data, acceleration and strain monitoring data of the in-service jacket platform under test; S2: Based on the actual structural parameters, a finite element model of the jacket platform is established, and the natural frequency and local strain of the jacket platform are obtained through simulation. Based on acceleration and strain monitoring data, the natural frequency and local strain of the jacket platform are obtained in real time. Using the errors corresponding to the simulated and real-time obtained natural frequency and local strain as variables, respectively, the target value of the finite element model correction is determined by constructing an objective function. By adjusting the model parameters of the finite element model, when the error of the target value meets the error threshold requirement, the corrected finite element model of the jacket platform is obtained. S3: Using real wind and wave load data as excitation, input the finite element modified model of the jacket platform to obtain the stress response dataset and displacement response dataset of the jacket platform through simulation; S4: Obtain the optimal modal basis vectors and their corresponding modal coefficients corresponding to the stress response dataset and displacement response dataset; use different preset optimization algorithms to predict the modal coefficients corresponding to the optimal modal basis vectors of different datasets respectively, and linearly combine and superimpose the obtained optimal modal basis vectors corresponding to different datasets with the predicted modal coefficients to construct stress response order reduction models and displacement response order reduction models respectively, and monitor the stress response and displacement response of the in-service jacket platform in real time.

[0020] The method proposed in this invention can dynamically predict changes in modal coefficients based on real wind and wave load data acquired at different times, and then update the reduced-order models of stress response and displacement response. This dynamic prediction capability enables the method to adapt to changes in the jacket platform under different working conditions and environmental conditions, and to predict its stress response and displacement response in a timely and accurate manner, providing a reliable basis for engineering decisions.

[0021] Specifically, in step S1, based on the design drawings of the in-service jacket platform, the actual structural parameters of the in-service jacket platform are collected, including material properties such as geometric dimensions, elastic modulus, density, and yield strength, as well as the connection methods between platform structures.

[0022] Real-time monitoring of the in-service jacket platform was carried out, and acceleration and strain data were collected by high-precision sensing equipment.

[0023] Real wind and wave load data includes time history data for both wind and wave loads, where: Based on the wind speed and wave sensors installed on the in-service jacket platform under test, wind speed and direction information are acquired in real time.

[0024] The time history data of wind load is obtained by constructing a wind pressure time history transformation based on the acquired wind speed and direction information, specifically: Based on the recommendations for wind load calculation in the DNVGL-RP-C205 specification, wind speed and direction information are converted into wind load components in the structural coordinate system through wind vector decomposition. Combined with parameters such as the shape and windward area of ​​the in-service jacket platform structure, the time history of wind loads acting on various parts of the structure is calculated.

[0025] Wind speed and direction information was collected using wind speed sensors installed on the in-service jacket platform. Wave load time history data was based on wave signal data. Using the Morison equation, wave forces were decomposed into inertial and drag forces, and the dominant wave frequency direction was simplified to a unidirectional dynamic input acting on the structure. Finally, all environmental excitation data were converted into standardized load time history data. Wave signal data was collected using wave sensors installed on the in-service jacket platform.

[0026] Step S1 ensures the temporal consistency and numerical stability of the acquired data, thereby meeting the requirements of subsequent model calculations.

[0027] In step S2, a finite element model of the jacket platform is constructed using the acquired actual structural parameters through the finite element analysis software ANSYS. Specifically, the pile legs, chord braces, and other structures in the finite element model are built using beam188 elements, the deck is built using shell181 elements, and the pipe nodes in the splash zone are built using soild187 elements. Multi-point constraint methods are used to bind the solid elements and beam elements.

[0028] Numerical simulation of the jacket platform was performed using finite element method to obtain the natural frequency and local strain of the in-service jacket platform.

[0029] Based on vibration monitoring experiments, acceleration and strain monitoring data were collected through sensing devices. The acceleration and strain data were preprocessed, including detrending, filtering and noise removal, to ensure the accuracy of subsequent analysis. Based on the preprocessed acceleration and strain data, the natural frequency and local strain of the in-service jacket platform were obtained.

[0030] Using the errors corresponding to the natural frequency and local strain obtained from simulation and real-time monitoring as variables, respectively, the target value for finite element model correction is determined by constructing an objective function; The objective function constructed is: ; In the formula: E This is the output of the objective function, i.e., the target value corrected by the finite element model; The natural frequency of the jacket platform is obtained through numerical simulation; To monitor and acquire the inherent frequency of in-service jacket platform in real time; These are the weighting coefficients; The local strain of the finite element model of the jacket platform obtained through numerical simulation; To monitor the local strain of the in-service jacket platform in real time.

[0031] The Kriging surrogate model combined with the Multi-Objective Genetic Algorithm (MOGA) is used as the model correction strategy for the finite element model. By adjusting the model parameters of the finite element model, including material property parameters such as material density, elastic modulus and pipe wall thickness, dimensional parameters and geometric feature parameters, the target value to be corrected is iteratively optimized until the error between the target value of the finite element model of the jacket platform and the natural frequency and local strain of the in-service jacket platform obtained by real-time monitoring meets the error threshold requirement, and the corrected finite element model of the jacket platform is obtained.

[0032] The Kriging surrogate model posits that the distribution of random variables is influenced by certain local biases on top of the global expectation, and that these local biases exhibit a Gaussian random distribution, with the degree of correlation represented by the covariance.

[0033] MOGA is a commonly used method for solving multi-objective optimization problems. This algorithm is a rewrite of the traditional genetic algorithm, which retains the basic processes of encoding, fitness function and selection, crossover and mutation of genetic operators, thus giving MOGA a good ability to solve nonlinear problems.

[0034] If the target value When the error is less than 1% of the error threshold, the correction requirement is met, and the finite element correction model of the jacket platform is obtained.

[0035] Steps S1-S2 comprehensively utilize multi-source information, including the actual structural parameters, acceleration, and strain monitoring data of the in-service jacket platform under test. A finite element model is established based on the actual structural parameters to obtain the natural frequencies and local strains. Simultaneously, these parameters are acquired in real-time using monitoring data. The error between these two models is used to construct an objective function for model correction. This multi-source data fusion approach can more comprehensively and accurately reflect the actual state of the jacket platform. Compared to correction methods based on a single data source, it can more effectively reduce the discrepancy between the finite element model and the actual structure, significantly improving the accuracy of the finite element model and laying a solid foundation for subsequent accurate stress and displacement response predictions.

[0036] The natural frequencies and local strains of the jacket platform are acquired in real time using acceleration and strain monitoring data, and the finite element model is corrected based on this. This real-time data-driven correction method can promptly reflect the structural changes and performance degradation of the jacket platform during actual use, ensuring that the model always maintains consistency with the actual structure and improving the model's real-time performance and adaptability.

[0037] In step S3, real wind and wave load data are used as excitation and input into the finite element modified model of the jacket platform. The stress response dataset and displacement response dataset of the finite element modified model of the jacket platform are obtained through simulation, specifically including: Real wind and wave load data includes time history data of wind load and wave load.

[0038] Based on the wind speed and wave sensors installed on the in-service jacket platform, wind speed and direction information are acquired in real time.

[0039] Based on wind speed and direction information, a wind pressure time history is constructed to obtain wind load time history data. The wind load time history data is then applied to the upper structural surface of the jacket platform finite element modified model.

[0040] Based on the Morison equation, time history data of wave loads are generated, and the time history data of wave loads are applied to the underwater part of the finite element modified model of the jacket platform in a node or element manner.

[0041] In the ANSYS environment, transient dynamic simulation analysis was performed on the finite element modified model of the jacket platform. The Newmark time integration method was used as the core algorithm of the solver for the finite element modified model of the jacket platform. The time step was set to 0.1s. The platform pipe nodes were taken as the research objects, and the stress response data and displacement response data simulated by the finite element modified model of the jacket platform were obtained.

[0042] Based on the stress response data and displacement response data simulated by the modified finite element model of the jacket platform, a stress response dataset is constructed. Q and displacement response dataset W .

[0043] This invention uses real wind and wave load data as excitation input into a modified finite element model to obtain stress response and displacement response datasets. Because it uses load data from the actual environment, it can more realistically simulate the stress conditions of the jacket platform under actual working conditions, making the obtained response data closer to reality and thus improving the accuracy of subsequent predictions.

[0044] In step S4, the stress response datasets are determined using the Proper Orthogonal Decomposition (POD) method. Q and displacement response dataset W The corresponding modal basis vectors and the modal coefficients corresponding to the modal basis vectors.

[0045] Based on stress response dataset Q and displacement response dataset WThe corresponding modal basis vectors and their corresponding modal coefficients are used to determine the stress response dataset based on leave-one-out cross-validation (LOOCV). Q and displacement response dataset W The corresponding optimal mode basis vector and its corresponding mode coefficients.

[0046] Based on the POD method, stress response datasets were obtained respectively. Q and displacement response dataset W The corresponding modal basis vectors and their corresponding modal coefficients are first obtained from the stress response dataset. Q and displacement response dataset W They are decomposed into three different matrices using the singular value decomposition method.

[0047] stress response dataset Q For example, it can be represented as: ; Where: matrix sum matrix All are unitary matrices, where the matrix Includes stress response dataset Q Correlation matrix The set of modal basis vectors; matrix It is a rectangular diagonal matrix, which is formed by All singular values ​​are arranged in descending order.

[0048] because With correlation matrix The non-zero eigenvalues ​​are all the same, so the correlation matrix can be solved first. eigenvalues and eigenvectors Therefore, we can conclude that: ; Therefore, the correlation matrix modal basis vectors It can be represented as: ; Therefore, the stress response of any node on the finite element model of the jacket platform can be represented by a linear combination: ; In the formula: For stress response dataset Q The modal coefficients corresponding to the modal basis vectors.

[0049] To construct a reduced-order model, it is necessary to select the optimal modal basis vectors and the modal coefficients corresponding to the optimal modal basis vectors, thereby realizing the stress response dataset. Q The reconstruction.

[0050] The optimal modal basis vectors are related to the selection of the modal order. A higher modal order results in a reduced-order model that retains system information closer to the finite element model, but it also increases the computational burden. Conversely, a lower modal order leads to insufficient accuracy in the reduced-order model. Therefore, the selection of the modal order needs to balance computational accuracy and efficiency.

[0051] However, traditional modal order truncation methods cannot determine whether the errors of reduced-order models constructed based on different modal orders converge on the same sample set. Therefore, this invention uses the LOOCV method to determine the optimal modal order selection criterion, thereby achieving optimized selection and determination of the modal order.

[0052] Based on stress response dataset Q The optimal modal basis vectors and their corresponding modal coefficients are obtained by combining the actual wind and wave load data with the previous data. n The modal coefficients corresponding to the optimal modal basis vectors are used as training data and input into an LSTM neural network. The Grey Wolf Optimizer (GWO) algorithm is used to optimize the hyperparameters of the LSTM neural network, including the number of hidden layer nodes, learning rate, and time step, to predict the stress response dataset. Q The modal coefficients corresponding to the optimal modal basis vectors; the stress response dataset Q The optimal modal basis vectors are linearly combined with the predicted modal coefficients and superimposed to construct a stress response order reduction model; Based on displacement response dataset W The optimal modal basis vectors and their corresponding modal coefficients will be used to combine the actual wind and wave load data with the previous data. m The modal coefficients corresponding to the optimal modal basis vectors are used as training data and input into a Radial Basis Function (RBF) neural network. The hyperparameters of the RBF neural network, including the number of hidden layer nodes, basis function width, and regularization coefficient, are optimized using the Sparrow Search Algorithm (SSA). This is used to predict the displacement response dataset. W The modal coefficients corresponding to the optimal modal basis vectors; the displacement response dataset W The optimal modal basis vectors and the predicted modal coefficients are linearly combined and superimposed to construct a displacement response reduced-order model.

[0053] The optimal modal basis vectors and their modal coefficients corresponding to the stress response and displacement response datasets are obtained. Based on the modal coefficients of different datasets, a pre-defined optimization algorithm is used to predict the modal coefficients corresponding to the optimal modal basis vectors. This modal analysis-based method can capture the main characteristics and dynamic properties of the jacket platform response. Predicting modal coefficients through optimization algorithms further improves the adaptability to response changes. Linearly combining and superimposing the optimal modal basis vectors and predicted modal coefficients to construct a reduced-order model can more accurately describe the stress and displacement responses while reducing computational complexity, thus enhancing the accuracy of the prediction results.

[0054] The constructed reduced-order model significantly reduces computational load by preserving the system's main dynamic characteristics and decreasing the model's degrees of freedom. In subsequent response prediction and analysis, using the reduced-order model can greatly shorten computation time and improve computational efficiency, making it particularly suitable for scenarios requiring frequent response prediction and real-time monitoring, thus meeting the demand for rapid response in practical engineering.

[0055] Based on real wind and wave load data acquired at different times, an optimization algorithm is used to dynamically predict changes in modal coefficients, thereby updating the reduced-order models of stress and displacement responses. This dynamic prediction capability enables the method to adapt to changes in the jacket platform under different working conditions and environmental conditions, accurately and timely predicting its stress and displacement responses, and providing a reliable basis for engineering decisions.

[0056] This invention also proposes a stress response and displacement response prediction system for a jacket platform based on LabVIEW and MATLAB, comprising: The data acquisition module is used to acquire the actual structural parameters, actual wind and wave load data, and acceleration and strain monitoring data of the in-service jacket platform under test. The actual wind and wave load data is acquired in real time by wind speed sensors and wave sensors (wave height meters) installed on the in-service jacket platform, while the acceleration and strain monitoring data are collected separately by high-precision sensing devices (accelerometers, etc.) installed on the in-service jacket platform.

[0057] The finite element model correction module is used to establish a finite element model of the jacket platform based on real structural parameters, and to obtain the natural frequency and local strain of the jacket platform through simulation. Based on acceleration and strain monitoring data, the natural frequency and local strain of the jacket platform are obtained in real time. Using the errors corresponding to the natural frequency and local strain obtained in simulation and real time as variables, respectively, the target value of the finite element model correction is determined by constructing an objective function. By adjusting the model parameters of the finite element model, when the error of the target value meets the error threshold requirement, the corrected finite element model of the jacket platform is obtained. The finite element simulation module is used to take real wind and wave load data as excitation, input it into the finite element modified model of the jacket platform, and obtain the stress response dataset and displacement response dataset of the jacket platform through simulation. The prediction module is used to obtain the optimal modal basis vectors and their corresponding modal coefficients corresponding to the stress response dataset and displacement response dataset. By predicting the modal coefficients corresponding to the optimal modal basis vectors of different datasets, the module linearly combines and superimposes the obtained optimal modal basis vectors of different datasets with the predicted modal coefficients to construct stress response reduction model and displacement response reduction model respectively, and monitors the stress response and displacement response of the in-service jacket platform in real time.

[0058] It also includes a dynamic response prediction module, which is used to input real wind and wave load data into the prediction module loaded on the land host to realize real-time prediction of stress and displacement response. The prediction module is loaded on the land host, and the real wind and wave load data is transmitted to the land host via the Beidou system. The results display and storage module visualizes and stores the stress and displacement responses output by the prediction module. Results generated by the prediction module are automatically recorded in the tables of the results display and storage module, ensuring that the numbers of each monitoring point and their corresponding predicted values ​​are completely preserved. This module not only supports real-time data storage but also facilitates users' querying and analysis of historical data at any time. Furthermore, users can export selected historical data to a designated folder for flexible later retrieval, based on actual needs.

[0059] The early warning module is used to determine whether there is a risk to the safety status of the jacket platform and to determine the degree of risk.

[0060] Stress warning: The stress response can be monitored in real time through the prediction module, and the stress response of key nodes can be evaluated online. The system uses the allowable stress as the early warning threshold for real-time online evaluation.

[0061] Displacement warning: Displacement response can be monitored in real time through the prediction module, enabling online evaluation of the platform's displacement response. When the first strut of the platform yields, local failure occurs, and the resulting displacement value is set as the first-level warning value. When the first pile leg of the platform yields, the structural bearing capacity decreases, and the resulting displacement value is set as the second-level warning value. When a node of the platform fails, it may cause collapse, and the displacement value at this time is set as the third-level warning value. When the system's displacement prediction results reach the corresponding warning threshold, an alarm is issued.

[0062] like Figure 2 As shown in this embodiment, a dynamic response online prediction system for a jacket platform based on a price reduction model is proposed. Figure 2The upper left figure shows the finite element modified model of the in-service jacket platform. A displacement monitoring point 1# is set on the top plate of the finite element modified model of the jacket platform; a stress monitoring point 1# is set on the key pipe node of the finite element modified model of the jacket platform. Figure 2 The upper right figure shows the environmental data collected in real time by the sensing device, which is then converted into environmental loads and input into the finite element correction model of the in-service jacket platform.

[0063] Figure 2 The lower left figure shows the stress response data predicted at stress monitoring point 1 using a reduced-order stress response model of the jacket platform. Figure 2 The lower right figure shows the displacement response data predicted at displacement monitoring point 1# using the reduced-order displacement response model of the jacket platform.

[0064] The stress and displacement response prediction method for the jacket platform proposed in this invention calculates the natural frequency and local strain (theoretical values) by establishing a finite element model, compares them with the natural frequency and local strain (measured values) of the jacket platform, quantifies the error of the finite element model, and corrects the target value through data-driven correction, so that the finite element model shifts from "idealized assumptions" to "adaptation to real working conditions", thereby improving the accuracy of stress / displacement response prediction.

[0065] By determining the optimal modal basis vectors and modal coefficients corresponding to the stress and displacement response data of the jacket platform, a mapping relationship between the modal coefficients and the load is established, enabling rapid prediction. This reduces the real-time response prediction time from minutes to seconds, meeting the needs of emergency decision-making in engineering, and ultimately achieving a balance between "high precision and high efficiency".

[0066] The method proposed in this invention can shift from "theoretical simulation" to "real-world operating condition driven", covering the entire life cycle of design and operation and maintenance, and providing quantitative decision support for the safe and economical operation of marine platforms.

[0067] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

[0068] Furthermore, unless otherwise stated, all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. All references to this specification are incorporated by way of citation to disclose and describe methods relating to those references. In the event of any conflict with any incorporated reference, the content of this specification shall prevail.

Claims

1. A method of predicting stress response and displacement response of a jacket platform, characterized by, The method comprises the following steps: obtaining real structure parameters, real wind and wave load data, acceleration and strain monitoring data of the in-service jacket platform to be tested; based on the real structure parameters, a finite element model of the jacket platform is established, and the natural frequency and local strain of the jacket platform are obtained through simulation; based on the acceleration and strain monitoring data, the natural frequency and local strain of the jacket platform are obtained in real time; taking the errors of the natural frequency and local strain obtained through simulation and in real time as variables, a target function is constructed to determine the target value of the finite element model correction; by adjusting the model parameters of the finite element model, when the error of the target value meets the error threshold requirement, a finite element correction model of the jacket platform is obtained; the real wind and wave load data are input into the finite element correction model of the jacket platform as excitation, and stress response data set and displacement response data set of the jacket platform are obtained through simulation; optimal modal basis vectors corresponding to the stress response data set and the displacement response data set and modal coefficients corresponding to the optimal modal basis vectors are obtained; the modal coefficients corresponding to the optimal modal basis vectors corresponding to different data sets are predicted, and the optimal modal basis vectors corresponding to different data sets and the predicted modal coefficients are linearly combined and superimposed to construct a stress response reduction model and a displacement response reduction model, respectively, to monitor the stress response and displacement response of the in-service jacket platform in real time.

2. The method of predicting the stress response and displacement response of a jacket platform according to claim 1, wherein, The optimal modal basis vectors corresponding to the stress response data set and the displacement response data set and the modal coefficients corresponding to the optimal modal basis vectors are obtained, and the optimal modal basis vectors corresponding to different data sets are predicted, and the optimal modal basis vectors corresponding to different data sets and the predicted modal coefficients are linearly combined and superimposed to construct a stress response reduction model and a displacement response reduction model, respectively, to monitor the stress response and displacement response of the in-service jacket platform in real time. Based on the proper orthogonal decomposition method, the stress response dataset Q and the displacement response dataset W corresponding modal basis vectors and modal coefficients corresponding to the modal basis vectors corresponding modal basis vectors and modal coefficients corresponding to the modal basis vectors, based on leave-one-out cross-validation, determine optimal modal basis vectors and modal coefficients corresponding to the optimal modal basis vectors, respectively, for the stress response dataset Q and the displacement response dataset W corresponding modal basis vectors and modal coefficients corresponding to the modal basis vectors, based on leave-one-out cross-validation, determine optimal modal basis vectors and modal coefficients corresponding to the optimal modal basis vectors, respectively, for the stress response dataset Q and the displacement response dataset W corresponding modal basis vectors and modal coefficients corresponding to the modal basis vectors, based on leave-one-out cross-validation, determine optimal modal basis vectors and modal coefficients corresponding to the optimal modal basis vectors, respectively, for the stress response dataset 3. The method of predicting the stress response and displacement response of a jacket platform according to claim 2, wherein, The real wind and wave load data are input into the finite element correction model of the jacket platform as excitation, and stress response data set and displacement response data set of the jacket platform are obtained through simulation, and the optimal modal basis vectors corresponding to the stress response data set and the displacement response data set and the modal coefficients corresponding to the optimal modal basis vectors are obtained. Based on stress response dataset Q The optimal modal basis vectors and their corresponding modal coefficients are obtained by combining the actual wind and wave load data with the previous data. n The modal coefficients corresponding to the optimal modal basis vectors are used as training data and input into an LSTM neural network. The Grey Wolf optimization algorithm is used to optimize the hyperparameters of the LSTM neural network, including the number of hidden layer nodes, learning rate, and time step, to predict the stress response dataset. Q The modal coefficients corresponding to the optimal modal basis vectors; the stress response dataset Q The optimal modal basis vectors are linearly combined with the predicted modal coefficients and superimposed to construct a stress response reduced-order model, and the stress response of the in-service jacket platform is monitored in real time. Based on displacement response dataset W The optimal modal basis vectors and their corresponding modal coefficients are obtained by combining the actual wind and wave load data with the previous data. m The modal coefficients corresponding to the optimal modal basis vectors are used as training data and input into the radial basis function neural network (RBN). The hyperparameters of the RBN, including the number of hidden layer nodes, basis function width, and regularization coefficient, are optimized using a sparrow search algorithm to predict the displacement response dataset. W The modal coefficients corresponding to the optimal modal basis vectors; the displacement response dataset W The optimal modal basis vectors and the predicted modal coefficients are linearly combined and superimposed to construct a displacement response reduced-order model, and the displacement response of the in-service jacket platform is monitored in real time.

4. The method of predicting the stress response and displacement response of a jacket platform according to claim 2, wherein, The real wind and wave load data include time history data of wind load and wave load; real-time wind speed and direction information is obtained based on wind speed sensors and wave sensors installed on the in-service jacket platform; based on the wind speed and direction information, a wind pressure time history is constructed, time history data of wind load is obtained, and the time history data of wind load is applied to the upper structure surface of the finite element correction model of the jacket platform; based on the Morison equation, time history data of wave load is generated, and the time history data of wave load is applied to the underwater structure of the finite element correction model of the jacket platform in the form of nodes or elements; in the ANSYS environment, transient dynamics simulation analysis is performed on the finite element correction model of the jacket platform, Newmark time integration method is used as the core algorithm of the solver of the finite element correction model of the jacket platform, the time step is set to 0.1s, the platform pipe nodes are taken as the research object, and the stress and displacement response data of the jacket platform are obtained. ​ Based on the acquired stress response data and displacement response data of the jacket platform, a stress response dataset is constructed Q and a displacement response dataset W .

5. The method of predicting the stress response and displacement response of a jacket platform according to claim 1, wherein, The real structure parameters are used to establish a finite element model of the jacket platform, and the natural frequency and local strain of the jacket platform are obtained through simulation, specifically including: The real structure parameters of the in-service jacket platform include material properties, size parameters and geometric characteristics; Based on the material properties, size parameters and geometric characteristics of the in-service jacket platform, a finite element model of the jacket platform is established by using ANSYS software, wherein the beam188 element is used to establish the pile leg and the string bar brace structure of the finite element model, the shell181 element is used to establish the deck of the finite element model, and the soild187 element is used to establish the pipe node of the splash zone of the deck. The finite element technology is used to simulate the jacket platform to obtain the natural frequency and local strain of the jacket platform.

6. The method of predicting the stress response and displacement response of a jacket platform according to claim 5, wherein, The errors corresponding to the natural frequency and local strain obtained through simulation and real-time acquisition are used as variables to determine the target value of the finite element model correction by constructing a target function, specifically including: According to the natural frequency and local strain obtained through simulation and real-time acquisition, the target function is constructed as: ； In the formula: E is the output of the objective function, i.e. the target value of the finite element model correction; is the natural frequency of the jacket platform obtained by numerical simulation; is the natural frequency of the in-service jacket platform obtained by real-time monitoring; is the weight coefficient; is the local strain of the finite element model of the jacket platform obtained by numerical simulation; is the local strain of the in-service jacket platform obtained by real-time monitoring; The output of the target function is used as the target value of the finite element model correction.

7. The method of predicting the stress response and displacement response of a jacket platform according to claim 6, wherein, The model parameters of the finite element model are adjusted, and when the error of the target value meets the error threshold requirement, the finite element correction model of the jacket platform is obtained, specifically including: The model correction strategy of the finite element model is obtained by combining the Kriging surrogate model with the multi-objective genetic algorithm. The model parameters of the finite element model, including the material density, elastic modulus and wall thickness of the pipe wall, are adjusted to iteratively optimize the target value to be corrected until the error between the target value of the finite element model of the jacket platform and the natural frequency and local strain of the in-service jacket platform obtained through real-time monitoring meets the error threshold requirement, and the finite element correction model of the jacket platform is obtained.

8. The method of predicting the stress response and displacement response of a jacket platform according to claim 1, wherein, The natural frequency and local strain of the jacket platform are obtained in real time based on the acceleration and strain monitoring data, specifically including: Acceleration and strain monitoring data are collected by a sensing device based on a vibration monitoring experiment; The acceleration and strain data are preprocessed, including detrending, filtering and noise removal; Based on the preprocessed acceleration and strain data, the natural frequency and local strain of the in-service jacket platform are obtained.

9. A system for predicting stress response and displacement response of a jacket platform, characterized by, It includes: A data acquisition module is used to acquire the real structure parameters, real wind and wave load data, acceleration and strain monitoring data of the in-service jacket platform to be tested; A finite element model correction module is used to establish a finite element model of the jacket platform based on the real structure parameters, and obtain the natural frequency and local strain of the jacket platform through simulation. The natural frequency and local strain of the jacket platform are obtained in real time based on the acceleration and strain monitoring data. The errors corresponding to the natural frequency and local strain obtained through simulation and real-time acquisition are used as variables to determine the target value of the finite element model correction by constructing a target function. The model parameters of the finite element model are adjusted, and when the error of the target value meets the error threshold requirement, the finite element correction model of the jacket platform is obtained. A finite element simulation module is configured to input real wave load data as excitation into a finite element correction model of the jacket platform to obtain a stress response data set and a displacement response data set of the jacket platform through simulation; A prediction module is configured to obtain optimal modal basis vectors corresponding to the stress response data set and the displacement response data set and modal coefficients corresponding to the optimal modal basis vectors; the prediction module is further configured to obtain the modal coefficients corresponding to the optimal modal basis vectors corresponding to different data sets through prediction, and linearly combine and superimpose the optimal modal basis vectors corresponding to different data sets and the predicted modal coefficients to respectively construct a stress response reduced-order model and a displacement response reduced-order model, so as to monitor the stress response and the displacement response of the in-service jacket platform in real time.

Images