Wharf and foundation integrated coupling modeling and simulation method considering large deformation and hysteretic characteristics of soil body

Abstract

The invention provides a wharf and foundation integrated coupling modeling and simulation method considering large deformation and hysteretic characteristics of a soil body, comprising the following steps: step 1, constructing a wharf and foundation integrated simulation analysis model for processing real data of a wharf foundation construction site and outputting characteristic parameters; 2, constructing a multi-modal data fusion network, and performing fusion correction on the feature parameters output in the step 1 according to real data; and step 3, establishing a wharf foundation design monitoring closed-loop control mechanism, and optimizing the wharf foundation integrated simulation analysis model and the multi-modal data fusion network in combination with the real data in the step 1 and the data fused and corrected in the step 2, thereby realizing wharf and foundation integrated coupling modeling and simulation considering the large deformation and hysteresis characteristics of the soil body.

Description

Technical Field

[0001] The present invention relates to a method for integrated coupling modeling and simulation of a wharf and a foundation, in particular to a method for integrated coupling modeling and simulation of a wharf and a foundation taking into account large deformation and hysteresis characteristics of soil. Background Art

[0002] This section merely provides background information related to the present disclosure and is not necessarily prior art.

[0003] As the most basic transportation infrastructure, docks are subject to the coupling effects of multiple complex environments. On the one hand, the thickness of the foundation cover of port docks often exceeds 30m, involving marine soft soil, silty sand, silty clay, etc. The large deformation of the deep soft soil foundation has a significant impact on the deformation of the dock superstructure. On the other hand, the high frequency of typhoons in offshore areas and the extreme environmental factors cause the dock structure system to bear complex loads far exceeding the design benchmark, further affecting the service performance of the dock.

[0004] The current wharf structure design specifications (such as JTS167-2018 and BS6349) adopt the "structure-foundation layered design" method, which regards the superstructure as a rigid body and simplifies the foundation into a Winkler foundation model, ignoring the dynamic coupling effect between piles and soil. Secondly, the conventional Mohr-Coulomb or Drucker-Prager model cannot characterize the creep and cyclic load hysteresis characteristics of deep soft soil, and does not consider the spatiotemporal coupling effects of multiple complex environmental factors. At the same time, traditional wharf-foundation design mostly relies on static finite element models and regular manual inspections, which makes it difficult to capture the dynamic response under the coupling of complex loads and environmental factors in real time, resulting in a large deviation between the model prediction and the actual working conditions. In addition, existing monitoring systems generally face technical bottlenecks such as insufficient data fusion capabilities, delayed model dynamic updates, and single optimization objectives.

[0005] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present disclosure, and therefore may include information that does not constitute prior art known to ordinary technicians in the field. Summary of the Invention

[0006] Purpose of the invention: The technical problem to be solved by the present invention is to address the deficiencies of the existing technology and provide an integrated coupling modeling and simulation method for a wharf and foundation that takes into account the large deformation and hysteresis characteristics of the soil.

[0007] In order to solve the above technical problems, the present invention discloses an integrated coupled modeling and simulation method for a wharf and foundation taking into account large deformation and hysteresis characteristics of the soil, comprising the following steps:

[0008] Step 1: Build an integrated simulation analysis model for the wharf foundation to process the real data of the wharf foundation construction site and output characteristic parameters;

[0009] Step 2: Build a multimodal data fusion network and perform fusion correction on the feature parameters output in step 1 based on real data;

[0010] Step 3: Establish a closed-loop control mechanism for monitoring the wharf foundation design. Combining the real data described in step 1 with the fused and corrected data from step 2, optimize the integrated simulation analysis model for the wharf foundation and the multimodal data fusion network to achieve integrated coupled modeling and simulation of the wharf and foundation that takes into account the large deformation and hysteresis characteristics of the soil.

[0011] Furthermore, the construction of the integrated simulation analysis model of the wharf foundation described in step 1 includes the following steps:

[0012] Step 1-1, obtaining input parameters of the integrated simulation analysis model of the wharf foundation based on the real data of the wharf foundation construction site;

[0013] Step 1-2: Based on the Darendeli viscoelastic constitutive model, an integrated simulation analysis model of the wharf foundation is established;

[0014] Steps 1-3: collect data from the wharf foundation construction site in real time, input it into the wharf foundation integrated simulation analysis model, and obtain output characteristic parameters.

[0015] Furthermore, the step 1-1 of obtaining the input parameters of the integrated simulation analysis model of the wharf foundation specifically includes:

[0016] Obtain the material parameters of the upper structure and the soil parameters of the lower foundation;

[0017] Material parameters of the pier superstructure, including: density, Young's modulus, and Poisson's ratio of concrete;

[0018] Soil parameters of the wharf subgrade, including:

[0019] By drilling and sampling at the wharf project site and conducting in-situ tests and indoor geotechnical tests, the density ρ, porosity e, moisture content w and Poisson's ratio u of the soil in the target area were obtained;

[0020] The dynamic triaxial test instrument is used to measure the stress and strain changes of representative soils under gradually increasing cyclic loads, and the maximum dynamic elastic modulus, reference stress, material attenuation coefficient and current dynamic stress are obtained.

[0021] Furthermore, the establishment of the integrated simulation analysis model of the wharf foundation described in steps 1-2 includes the following steps:

[0022] Step 1-2-1: Establish a preliminary integrated simulation analysis model for the wharf foundation based on the Darendeli viscoelastic constitutive model. The input parameters are the maximum dynamic elastic modulus, reference stress, material attenuation coefficient, and current dynamic stress of the soil mechanical parameters. The output parameters are the stress at the current integration point, the material Jacobian matrix, and the state variable array.

[0023] Step 1-2-2: Based on the preliminary integrated simulation analysis model of the wharf foundation and the actual engineering background, an overall multi-scale model of the wharf foundation is established, i.e., the final integrated simulation analysis model of the wharf foundation. Specifically, it includes:

[0024] Step 1-2-2-1: Upper pier structure: construct the upper structure model and assign material properties, including density, Young's modulus, and Poisson's ratio of concrete;

[0025] Step 1-2-2-2, pile foundation system: the upper wharf structure and the lower foundation are connected through the pile foundation system. The equivalent shear layer model is adopted, and the softening effect of the pile side friction resistance is simulated through zero-thickness cohesive elements.

[0026] Step 1-2-2-3: For the lower soft soil foundation, assign soil parameters to foundation attributes, build a three-dimensional solid element model, and use a transition mesh in the area around the pile;

[0027] Step 1-2-2-3, boundary condition setting, the upper pier structure adopts free field boundary, and the lower soft soil foundation applies viscoelastic boundary condition;

[0028] Step 1-2-3, set the model loading strategy:

[0029] A graded loading strategy was adopted to gradually apply the actual on-site load data to the integrated simulation analysis model of the wharf foundation to simulate the consolidation process and nonlinear response;

[0030] The birth-death element method is used to pre-treat the foundation consolidation.

[0031] Furthermore, the construction of the multimodal data fusion network described in step 2 includes the following steps:

[0032] Step 2-1, pre-processing the characteristic parameters output by the integrated simulation analysis model of the wharf foundation;

[0033] Step 2-2: Build a multimodal data fusion network based on the Transformer network;

[0034] Step 2-3: Dynamically calibrate and optimize the multimodal data fusion network.

[0035] Furthermore, the characteristic parameters output by the integrated simulation analysis model of the wharf foundation as described in step 2-1 are pre-processed, including:

[0036] Wavelet transform is used to perform multi-scale decomposition processing on the characteristic parameters output by the integrated simulation analysis model of the wharf foundation and the real data obtained on site, as follows:

[0037] Assume that the input data is x(t), and perform continuous wavelet transform on it as follows:

[0038]

[0039] Among them, a is the scale parameter, b is the translation parameter, ψ is the selected wavelet function, by selecting the scale range a∈[a min ,a max ] and reconstruction methods to filter out high-frequency noise and extract key feature parameters.

[0040] Furthermore, step 2-2 constructs a multimodal data fusion network, including the following steps:

[0041] Step 2-2-1: Flatten the key feature parameter data of the spatial distribution to form a vector representation, specifically including:

[0042] The embedding layer is used to encode the data sources of different key feature parameters to achieve unified dimensional mapping of the data;

[0043] The self-attention module of the Transformer network is used to capture the correlation between time and space, compare the real-time monitoring data with key feature parameters, and extract key error information and structural response features;

[0044] Step 2-2-2: Improve feature expression capabilities through residual connections and multi-head attention mechanisms;

[0045] Step 2-2-3, optimize feature extraction and data enhancement, that is, introduce a feature extraction optimization module into the Transformer network to perform feature dimensionality reduction and information reconstruction on the fused multimodal data.

[0046] Furthermore, the dynamic correction and optimization of the multimodal data fusion network described in steps 2-3 includes the following steps:

[0047] Step 2-3-1, prediction error acquisition and correction;

[0048] Step 2-3-2, objective function and parameter tuning.

[0049] Furthermore, the prediction error acquisition and correction described in step 2-3-1 are as follows:

[0050] The real-time monitoring data is compared with the characteristic parameters output by the integrated simulation analysis model of the wharf foundation to obtain the prediction error between the two. Assuming that the characteristic parameter output by the integrated simulation analysis model of the wharf foundation is y and the actual monitoring value is z, the state prediction process adopts the linear dynamic system model as follows:

[0051] z m∣m-1 =Ayz m-1∣m-1 +Bl m-1

[0052] P m∣m-1 =AP k-1∣k-1 A T +Q

[0053] Among them, A is the state transfer matrix, B is the control matrix, l m-1 is the control input, Q is the process noise covariance matrix, P m∣m-1 is the prediction covariance matrix, z m∣m-1 is the prior prediction value of the monitoring value before time m, m represents the time of the mth prediction or observation, and k represents the time of the kth prediction or observation;

[0054] The prediction error correction amount is calculated by the above-mentioned Kalman filter recursive formula, and the internal parameters of the model are dynamically adjusted so that the corrected prediction value is closest to the monitoring value.

[0055] Furthermore, the objective function and parameter tuning described in step 2-3-2 specifically include:

[0056] Construct the objective function and set the prediction output of the integrated simulation analysis model of the wharf foundation as f(x i ,θ), the actual output is o, then the objective function is defined as:

[0057]

[0058] By solving the above least squares problem, the parameter θ of the integrated simulation analysis model of the wharf foundation is updated to minimize the prediction error. Combined with the Bayesian updating theory, historical data and real-time data are used to continuously update the prior distribution and continuously correct the model parameter distribution.

[0059] Beneficial effects:

[0060] 1. The method proposed in this invention can not only fully reflect the nonlinear response of soil under cyclic loads, but also simulate the complex coupling between the various components of the wharf.

[0061] 2. The method proposed in this invention realizes the real and dynamic coupling response simulation between the soft soil foundation and the wharf structure in a complex marine environment, thereby improving the accuracy and efficiency of engineering design and safety assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, and the above and / or other advantages of the present invention will become more apparent.

[0063] Figure 1 It is a schematic diagram of the overall process of constructing an integrated simulation analysis model for a wharf foundation according to the present invention.

[0064] Figure 2 It is a schematic diagram of the mechanical behavior of soil under cyclic stress in the present invention.

[0065] Figure 3 It is a schematic diagram of the effect of the overall multi-scale simulation model of the wharf foundation in the present invention.

[0066] Figure 4 It is a schematic diagram of the process of constructing the overall multi-scale simulation model of the wharf foundation in the present invention.

[0067] Figure 5 It is a schematic diagram of the overall architecture of the present invention.

[0068] Figure 6 The figure is a stress distribution cloud diagram of the wharf structure in an embodiment.

[0069] Figure 7 The figure is a cloud diagram of the settlement and displacement of the wharf foundation soil in an embodiment.

[0070] Figure 8 The figure is a cloud diagram of the settlement and displacement of a wharf structure under different load distribution modes in an embodiment. DETAILED DESCRIPTION

[0071] The overall concept of this invention is as follows: It proposes developing a viscoelastic constitutive subroutine flow, constructing an integrated coupled simulation model for the wharf, and building a closed-loop control system integrating high-precision monitoring, multimodal data fusion, and dynamic model correction. Taking into account the large deformation and hysteretic properties of the soil, this integrated model analyzes the mechanical response of the wharf under deep soft soil and complex loading conditions. The goal is to provide a highly integrated, engineered, and highly accurate simulation system that fully reflects the nonlinear response of the soil under cyclic loading and simulates the complex coupling interactions between the various wharf components. Furthermore, a Transformer network is introduced to analyze the spatiotemporal correlations of cross-modal data. The finite element model is dynamically modified using a Kalman filter and Bayesian updating theory. A multi-objective genetic algorithm is then used to generate a Pareto optimal solution set, ultimately forming a "monitoring-analysis-optimization" closed loop. This system achieves realistic and dynamic simulation of the coupled response between the soft soil foundation and the wharf structure in complex marine environments, improving the accuracy and efficiency of engineering design and safety assessment, and providing technical support for the full lifecycle management of wharf projects.

[0072] The method provided by the present invention, such as Figure 5 As shown, the following steps are included:

[0073] 1. Build a terminal integration simulation analysis model

[0074] 1.1 Obtaining input parameters of the integrated simulation model of the wharf foundation

[0075] The input parameters of the wharf simulation model include the material parameters of the superstructure and the soil parameters of the lower foundation.

[0076] The superstructure of the wharf is mainly made of concrete, which is obtained through relevant construction design drawings and material lists, mainly including: concrete density, Young's modulus, and Poisson's ratio.

[0077] For the lower foundation of the wharf, first, by drilling samples at the wharf project site and conducting in-situ tests and indoor geotechnical tests in accordance with the "Standard for Geotechnical Test Methods", the basic physical indicators of the soil in the target area, such as density ρ, porosity e, moisture content w, and Poisson's ratio u, were obtained. Secondly, the stress-strain (σ-ε) changes of representative soils under step-by-step cyclic loads were measured using a dynamic triaxial test instrument. The third hysteresis loop data of each loading level was selected, and the vertices of each hysteresis loop were connected to obtain the soil backbone curve. In view of the large deformation, hysteretic response, and energy dissipation characteristics of thick soils under cyclic loads, the Darendeli viscoelastic constitutive model was adopted, and its mathematical expression is:

[0078]

[0079] Among them, E max is the maximum dynamic elastic modulus; σ r is the reference stress, which is when E / E max = 0.5; c is the material attenuation coefficient; σ is the current dynamic stress; E max According to the dynamic elastic modulus value when ε approaches 0 in the Hardin model, j and k are fitting parameters, and the formula is as follows:

[0080]

[0081] Therefore, the main soil foundation input parameters that can be obtained are: ρ, e, w, E max , σ r ,c,u.

[0082] 1.2 Establishment of a pier-foundation integrated simulation model based on soil viscoelastic constitutive model

[0083] First, based on the above Darendeli viscoelastic constitutive model, the corresponding user subroutine is written and embedded into the simulation analysis. The input parameters in the subroutine are the soil mechanical parameters E max , σr , c, u, and the output parameters are STRESS(NTENS), DDSDDE(NTENS,NTENS), and STATEV(NSTATV). STRESS(NTENS) is the stress at the current integration point; DDSDDE(NTENS,NTENS) is the material Jacobian matrix (tangent stiffness matrix); and STATEV(NSTATV) is the state variable array used to store historical variables in the material model. The key steps are as follows:

[0084] (1) Declare the subroutine interface and define the input and output parameters.

[0085] (2) Read and verify material parameters. Ensure that the number and type of input parameters match, and initialize the soil physical properties.

[0086] (3) Stress decomposition and invariant calculation to obtain equivalent dynamic stress σ d First, calculate the mean stress σ m and deviatoric stress s ij ,σ m Represents the isotropic pressure component of the material.

[0087]

[0088] Among them, σ 11 ,σ 22 ,σ 33 is the normal stress component.

[0089] Secondly, the J2 invariant is obtained by the following formula and converted into the equivalent dynamic stress. The specific formula is as follows.

[0090]

[0091] The equivalent dynamic stress σ d This is STRESS (NTENS), one of the output quantities.

[0092] (4) Substitute into Darendeli formula to calculate E. According to the current equivalent dynamic stress σ d Calculate and update the dynamic elastic modulus E.

[0093] (5) Establish the elastic matrix. First, calculate the elastic matrix to obtain the shear modulus G, bulk modulus K, and Lamé constant λ, then initialize the Jacobian matrix, and finally fill in the elastic matrix. The formula is as follows

[0094]

[0095] Through this step, the key parameters G, K and λ of DDSDDE (NTENS, NTENS) can be obtained.

[0096] (6) Update stress and record the corresponding historical variables, which is one of the output variables STATEV (NSTATV)

[0097] Secondly, based on the actual engineering background, a multi-scale model of the wharf-foundation is established. The model includes:

[0098] (1) Upper wharf structure: The upper structure model is carefully constructed and material properties are assigned, including the density, Young's modulus, and Poisson's ratio of the concrete.

[0099] (2) Pile foundation system. The upper structure and the lower foundation are connected through pile foundations. In the pile foundation area, the “equivalent shear layer” model is adopted to simulate the softening effect of the pile side friction resistance through zero-thickness cohesive units to achieve a true reflection of the interaction between the pile and the soil.

[0100] (3) Lower soft soil foundation. The soil parameters obtained above (density ρ, void ratio e, water content w, Poisson's ratio u) are assigned to the foundation attributes, and a three-dimensional solid unit model is constructed. The transition mesh technology is used in the area around the pile (the range is controlled within 5 times the pile diameter) to ensure that the local mesh size is no larger than 0.5 m to improve the accuracy of the numerical calculation.

[0101] (4) Boundary condition setting: The upper structure adopts a free field boundary to eliminate boundary reflection; the lower foundation applies viscoelastic boundary conditions to simulate the actual infinite domain environment.

[0102] Furthermore, set the model loading strategy

[0103] During the design and analysis process, a graded loading strategy was adopted, gradually applying actual on-site loads to simulate the consolidation process and nonlinear response. Simultaneously, the birth-death element method was used to pre-treat the foundation. To address local softening and large deformation issues, elements were dynamically deleted or restored to more realistically simulate the soil consolidation and instability process.

[0104] 1.3 Output characteristic parameters based on the terminal-foundation integrated simulation model

[0105] Based on the aforementioned wharf model, various environmental parameters (such as yard load and soil density) and structural parameters (such as concrete strength, pile layout, and pile diameter) were varied to analyze the settlement, deformation, and stability of the wharf foundation under different operating conditions. Changes in strain, displacement, and energy of the overall structure were quantitatively described. Combined with on-site monitoring, a variety of sensors (such as settlement sensors, strain gauges, and accelerometers) were deployed at key locations on the wharf and foundation to collect real-time structural physical data. This enabled accurate prediction of the wharf-foundation coupling response and the distribution of local vulnerable points, enabling real-time updates of the model state and forming a highly accurate coupled simulation system. Finally, key characteristic parameters were extracted, including stress and strain values at local vulnerable points on the wharf and displacement and settlement of the wharf cap.

[0106] 2. Building a multimodal data fusion network

[0107] By deeply integrating structured sensor data (settlement, strain, vibration) with unstructured visual data (images / video) and model predictions, a multimodal data fusion network enables refined perception and adaptive correction of the full lifecycle status of the terminal and foundation. Its advantages primarily lie in improved perception coverage, enhanced robustness, and enhanced prediction accuracy and real-time performance. Ultimately, this brings significant economic and technical benefits for structural health assessment, early warning response, and operational optimization.

[0108] The input parameters of the multimodal data fusion network are structured sensor data (such as settlement, strain, vibration data), unstructured data (such as on-site images, videos, etc.), historical record data and simulation model data.

[0109] The output parameter of the model is the fused feature vector F. The Transformer network performs multi-head self-attention and residual connections on multi-source inputs to obtain a unified high-dimensional representation. This vector contains both the spatiotemporal correlation information of real-time monitoring and visual features, as well as key signals such as prediction errors, which are used for subsequent model updates and optimization processes.

[0110] 2.1 Data Preprocessing

[0111] Wavelet transform is used to perform multi-scale decomposition on the above simulation model and the data obtained from field monitoring to effectively separate environmental noise, interference signals and model errors, and retain the true monitoring signal. Let the signal be x(t), and the formula for its continuous wavelet transform (CWT) is as follows:

[0112]

[0113] Among them, a is the scale parameter, b is the translation parameter, and ψ is the selected wavelet function. By choosing a suitable scale range a∈[a min ,a max ] and reconstruction methods to filter out high-frequency noise and extract representative feature parameters.

[0114] Key characteristic parameters such as the amplitude of local strain change, sedimentation rate change, and spectral characteristics of the vibration signal are extracted through wavelet packet decomposition and reconstruction, providing accurate input information for subsequent data fusion and model correction.

[0115] 2.2 Building a multimodal data fusion network

[0116] 2.2.1 Multi-source data fusion requirements

[0117] Monitoring data comes from a variety of sources, including structured sensor data (such as settlement, strain, and vibration data), unstructured data (such as on-site images and videos), historical records, and simulation model data. These data differ in time, space, and data type. Therefore, one of the key technologies in this system is to use the Transformer network, a deep learning framework with powerful sequence modeling and attention mechanisms, to perform multimodal data fusion.

[0118] 2.2.2 Transformer Fusion Network Design

[0119] First, the spatial distribution data is flattened to form a vector representation:

[0120] Use embedding layers to encode features of different data sources and achieve unified dimensional mapping of data;

[0121] The Transformer's self-attention module is used to capture the correlation between time and space, compare the real-time monitoring data with the pre-established finite element model prediction data, and extract key error information and structural response characteristics;

[0122] At the same time, the residual connection and multi-head attention mechanism are used to improve the feature expression ability, so that the data after information fusion has higher representativeness and distinguishing ability. Among them, the typical self-attention calculation formula is:

[0123]

[0124] Here, H (query vector), W (key vector), and V (value vector) are all obtained by passing the input data through different embedding layers, and d is the dimension of the key vector. This mechanism can capture the spatiotemporal correlations and key features between data, and the output fused feature vector F provides high-quality input for subsequent model updates.

[0125] 2.2.3 Optimizing feature extraction and data enhancement

[0126] In the Transformer network architecture, a feature extraction and optimization module is introduced to further perform feature dimensionality reduction and information reconstruction on the fused multimodal data, thereby improving the expressiveness and information utilization of the dataset and providing high-precision input for subsequent model status updates.

[0127] 2.3 Model updating and dynamic correction

[0128] 2.3.1 Principles of Prediction Error Acquisition and Correction

[0129] The real-time monitoring data is compared with the prediction results of the pre-established integrated wharf-foundation analysis model to obtain the prediction error between the two. Assuming that the state parameter predicted by the finite element model is y (for example, predicted settlement, stress value, etc.), and the actual monitoring value is z, the state prediction process can be based on a linear dynamic system model:

[0130] z m∣m-1 =Ayz m-1∣m-1 +Bl m-1

[0131] P m∣m-1 =AP k-1∣k-1 A T +Q

[0132] Among them, A is the state transfer matrix, B is the control matrix, l m-1 is the control input, Q is the process noise covariance matrix, P is the prediction covariance matrix, m represents the time of the mth prediction or observation, and k is similar to m in some formulas and also represents the general symbol in the time step.

[0133] The Kalman filter recursive formula is used to calculate the prediction error correction, dynamically adjusting the model's internal parameters to ensure that the corrected prediction value is closest to the monitored value. The specific steps include setting the system state vector and its prior distribution; and updating the system state estimate based on actual monitoring data.

[0134] 2.3.2 Objective Function and Parameter Tuning

[0135] Construct an objective function that minimizes the deviation between the model prediction value and the actual monitoring data through the least squares method, and converts the prediction error into an optimization index. Let the model prediction output be f(x i ;θ), the actual output is o, then the objective function is defined as:

[0136]

[0137] By solving the above least squares problem, the model parameters θ are updated to minimize the prediction error. Combined with the Bayesian updating theory, historical data and real-time data are used to continuously update the prior distribution and continuously correct the model parameter distribution to achieve the purpose of improving the model prediction accuracy.

[0138] 3. Establish a closed-loop control method for dock foundation design monitoring

[0139] 3.1 Automatic monitoring and feedback triggering mechanism

[0140] Integrate the above-mentioned monitoring data collection, data fusion, and model optimization modules into an automatic closed-loop system platform, including:

[0141] Data processing center: realizes real-time data acquisition, wavelet transform filtering and feature extraction;

[0142] Deep learning operation module: multi-source data fusion based on Transformer;

[0143] Dynamic model update module: integrates Kalman filtering, least squares parameter adjustment and Bayesian update algorithm;

[0144] Optimization algorithm server: implements multi-objective genetic algorithm optimization;

[0145] Visualization and early warning terminal: displays monitoring data, model predictions, and optimization results in real time, and triggers an early warning mechanism. When the deviation between the real-time monitoring data z and the model prediction value θ is greater than the preset tolerance δ, the system automatically executes the "update parameters and re-optimize" operation, enabling the closed-loop control system to capture structural state changes in real time and use the latest feedback to update the model and output the optimized design solution.

[0146] In summary, the entire process of data collection → analysis and processing → feedback correction is automated. The system sets a monitoring threshold. When the real-time monitoring data exceeds the set value, the feedback mechanism is automatically triggered to activate the model update and solution optimization modules.

[0147] 3.2 Platform Integration and Full Life Cycle Management

[0148] The integrated platform includes a data processing center, a deep learning computing module, an optimization algorithm server, and a visualization terminal. The system not only outputs real-time information on the wharf and foundation settlement, displacement, stress, and vibration, but also provides optimized design parameters and operational recommendations to the project's operations and maintenance department. This enables dynamic design optimization and risk warning throughout the entire lifecycle, improving project safety and cost-effectiveness.

[0149] 3.3 Closed-loop control effect

[0150] Through the "Monitor → Analysis → Feedback" function, the system can capture changes in structural status in real time, dynamically adjust the finite element model and optimization algorithm parameters, and ensure that the predicted model closely matches the actual operating conditions. Ultimately, the system automatically outputs the optimized design plan and operation and maintenance warnings, achieving an organic balance between structural safety, service performance, and economic efficiency. This system establishes a closed-loop control system for the overall design and service performance optimization of the wharf and foundation, "Monitoring Data → Model Update → Solution Optimization," providing a scientific basis for engineering decision-making.

[0151] Example:

[0152] The embodiment of the present invention is a bulk terminal in Zhanjiang. In this embodiment, Figure 1 As shown in the figure, a model method considering large deformation and hysteresis characteristics of soil is provided, which includes:

[0153] First, drilling and sampling of the soil in the target area was carried out at the wharf project site. According to the "Standard for Geotechnical Test Methods", the basic physical indicators of the soil were measured through on-site in-situ tests and indoor geotechnical tests, including basic physical parameters such as soil density ρ, porosity e, moisture content w, and Poisson's ratio u. The stress-strain (σ-ε) change value of the representative soil under cyclic loading was measured using a dynamic triaxial test instrument. The dynamic triaxial test adopts a loading mode of gradually increasing cyclic loads. The waveform is selected as a sine wave, the loading frequency is 1Hz, and each level of cyclic loading is 5 weeks. 100 points are collected and recorded each week. The third circle of each level of loading is selected as the test data. The backbone curve of the soil, that is, the stress-strain curve, is obtained by connecting the vertices of each hysteresis loop, as shown in the figure. Figure 2 As shown. In view of the large deformation and hysteresis characteristics of thick soft soil under cyclic loading, in order to ensure that the model can reflect the stiffness attenuation and energy dissipation characteristics of soft soil under large strain conditions, the Darendeli viscoelastic constitutive model is used, and its mathematical expression is:

[0154]

[0155] In, E max is the maximum dynamic elastic modulus; σ r is the reference stress, which is when E / E max = 0.5; c is the material attenuation coefficient; σ is the current dynamic stress (usually the square root of the second invariant of the deviatoric stress J2); E max From the Hardin model, when ε d The dynamic elastic modulus value when it approaches 0 is as follows:

[0156]

[0157] Where a and b are fitting parameters.

[0158] The above method can be used to obtain the model input parameters of a bulk terminal in Zhanjiang as shown in Table 1:

[0159] Table 1 Model input parameters for a bulk terminal in Zhanjiang

[0160]

[0161] Based on the Darendeli viscoelastic constitutive model, a subroutine was written and embedded in the iterative finite element model of the terminal. This example uses ABAQUS finite element simulation software. The steps, specific code, and comments for writing the subroutine are as follows:

[0162] (1) Declare the subroutine interface and define the input and output parameters

[0163]

[0164]

[0165]

[0166]

[0167]

[0168] The accuracy of the bulk viscoelastic constitutive model was further verified. The correctness of the subroutine was verified in simple models (e.g., uniaxial compression models), and the computational stability and convergence were checked. A simple uniaxial compression simulation of the soil was performed, and the modulus decay was observed by applying an axial displacement to the soil. The calculated stress-strain curve was then verified to be consistent with the theoretical value of the Darendeli model. The E-σ curve output by the model was also verified to be consistent with the variation pattern observed in indoor triaxial tests.

[0169] In this embodiment, a high-precision wharf-foundation integrated model is as follows Figure 3 As shown, it includes the upper yard structure of the wharf, cap, pier foundation, vertical piles, inclined piles and soft soil foundation. Figure 4 This is a flow chart of a high-precision wharf-foundation integrated coupling simulation system.

[0170] According to the actual engineering background, the overall multi-scale model of the wharf foundation was established, and the wharf superstructure model was refined. The elastic-plastic damage model was used to describe the stress response characteristics of concrete, and the physical and mechanical parameters of the concrete material were assigned, including a density of 2.55g / cm 3, Young's modulus is 34.58GPa, and Poisson's ratio is 0.2. Secondly, the physical and mechanical parameters of the soil obtained from the in-situ test and the indoor test are input into the material properties to construct a three-dimensional solid model of the foundation. The upper structure and the lower foundation are connected through the pile foundation to ensure the synergistic effect of the two parts in stress and deformation. In the pile foundation area, by establishing an "equivalent shear layer" model, the zero-thickness cohesive unit is used to simulate the softening effect of the pile side friction resistance to reflect the interaction between the pile and the soil. The upper structure adopts a free field boundary to eliminate the boundary reflection effect; the viscoelastic boundary condition is applied to the lower foundation, and the transition grid technology is used to divide the grid to ensure that the grid size in the area around the pile (within 5 times the pile diameter) is controlled at ≤0.5m to improve the local calculation accuracy; the overall model unit adopts a three-dimensional solid unit. During the design and analysis steps, a graded loading strategy was adopted, gradually applying actual on-site loads. This enabled the model to simulate the consolidation process and nonlinear response. The subroutines developed in the first aspect (implementing the Darendeli viscoelastic constitutive calculation and elastic matrix update process) were embedded in the overall iterative model. Furthermore, the birth-death element method was used to pre-treat the foundation for consolidation. To address local soil softening and large deformation issues, elements were dynamically deleted or restored to simulate the actual soil consolidation process.

[0171] During the design and analysis process, a graded loading strategy was adopted, gradually applying actual on-site loads to simulate the consolidation process and nonlinear response. Simultaneously, the birth-death element method was used to pre-treat the foundation. To address local softening and large deformation issues, elements were dynamically deleted or restored to more realistically simulate the soil consolidation and instability process.

[0172] By changing different environmental parameters (such as yard load, soil density, etc.) and structural parameters (such as concrete strength, pile layout, pile diameter, etc.), the settlement, deformation and stability of the foundation under the yard load are simulated. The strain, displacement and energy changes of the overall structure under different load and environmental conditions are analyzed, and the synergy between the wharf structure and foundation and the distribution of local vulnerable points are quantitatively described. Combined with on-site monitoring data, the model parameters are dynamically adjusted to achieve accurate prediction of the structure-foundation coupling response. Secondly, a visual simulation environment is constructed based on professional simulation software to achieve real-time output of the wharf operation process and foundation deformation, synchronously integrate historical monitoring data and real-time monitoring data, evaluate the foundation stability, and use data feedback to update the model status in real time, ultimately forming a high-precision wharf-foundation integrated coupling simulation system. Figure 6 、 Figure 7 and Figure 8 This is part of the simulation model output in this example. Figure 6 The figure is a stress distribution cloud diagram of the wharf structure in an embodiment. Figure 7 The figure is a cloud diagram of the settlement and displacement of the wharf foundation soil in an embodiment. Figure 8The figure is a cloud diagram of the settlement and displacement of a wharf structure under different load distribution modes in an embodiment.

[0173] In a specific implementation, the present application provides a computer storage medium and a corresponding data processing unit, wherein the computer storage medium is capable of storing a computer program that, when executed by the data processing unit, executes the invention content of the method for integrated coupled modeling and simulation of a wharf and foundation that considers large soil deformation and hysteresis characteristics, as well as some or all of the steps in each embodiment. The storage medium may be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0174] Those skilled in the art will clearly understand that the technical solutions in the embodiments of the present invention can be implemented by means of computer programs and their corresponding general hardware platforms. Based on this understanding, the technical solutions in the embodiments of the present invention, in essence or in other words, the part that contributes to the prior art, can be embodied in the form of a computer program, i.e., a software product. The computer program software product can be stored in a storage medium and includes a number of instructions for enabling a device including a data processing unit (which can be a personal computer, server, single-chip microcomputer, MCU, or network device, etc.) to execute the methods described in various embodiments of the present invention or certain parts of the embodiments.

[0175] This invention provides a method and approach for integrated coupled modeling and simulation of a wharf and foundation that considers large soil deformation and hysteretic characteristics. While numerous methods and approaches exist for implementing this technical solution, the aforementioned are merely preferred embodiments of the invention. It should be noted that those skilled in the art could readily make improvements and modifications without departing from the principles of the invention, and such improvements and modifications are considered within the scope of protection of this invention. Components not specified in this embodiment may be implemented using existing technologies.

Claims

1. A method for integrated coupled modeling and simulation of a wharf and foundation considering large deformation and hysteresis characteristics of soil, characterized by: The following steps are involved: Step 1: Build an integrated simulation analysis model for the wharf foundation to process the real data of the wharf foundation construction site and output characteristic parameters; Step 2: Build a multimodal data fusion network and perform fusion correction on the feature parameters output in step 1 based on real data; Step 3: Establish a closed-loop control mechanism for monitoring the wharf foundation design. Combining the real data described in step 1 with the fused and corrected data from step 2, optimize the integrated simulation analysis model for the wharf foundation and the multimodal data fusion network to achieve integrated coupled modeling and simulation of the wharf and foundation that takes into account the large deformation and hysteresis characteristics of the soil.

2. The method for integrated coupled modeling and simulation of a wharf and foundation considering large deformation and hysteresis characteristics of soil according to claim 1 is characterized in that: The construction of the integrated simulation analysis model of the wharf foundation described in step 1 includes the following steps: Step 1-1, obtaining input parameters of the integrated simulation analysis model of the wharf foundation based on the real data of the wharf foundation construction site; Step 1-2: Based on the Darendeli viscoelastic constitutive model, an integrated simulation analysis model of the wharf foundation is established; Steps 1-3: collect data from the wharf foundation construction site in real time, input it into the wharf foundation integrated simulation analysis model, and obtain output characteristic parameters.

3. The method for integrated coupled modeling and simulation of a wharf and foundation considering large deformation and hysteresis characteristics of soil according to claim 2 is characterized in that: The input parameters of the integrated simulation analysis model for the wharf foundation described in step 1-1 specifically include: Obtain the material parameters of the upper structure and the soil parameters of the lower foundation; Material parameters of the pier superstructure, including: density, Young's modulus, and Poisson's ratio of concrete; Soil parameters of the wharf subgrade, including: By drilling and sampling at the wharf project site and conducting in-situ tests and indoor geotechnical tests, the density ρ, porosity e, moisture content w and Poisson's ratio u of the soil in the target area were obtained; The dynamic triaxial test instrument is used to measure the stress and strain changes of representative soils under gradually increasing cyclic loads, and the maximum dynamic elastic modulus, reference stress, material attenuation coefficient and current dynamic stress are obtained.

4. The method for integrated coupled modeling and simulation of a wharf and foundation considering large deformation and hysteresis characteristics of soil according to claim 3 is characterized in that: The establishment of the integrated simulation analysis model of the wharf foundation described in steps 1-2 includes the following steps: Step 1-2-1: Establish a preliminary integrated simulation analysis model for the wharf foundation based on the Darendeli viscoelastic constitutive model. The input parameters are the maximum dynamic elastic modulus, reference stress, material attenuation coefficient, and current dynamic stress of the soil mechanical parameters. The output parameters are the stress at the current integration point, the material Jacobian matrix, and the state variable array. Step 1-2-2: Based on the preliminary integrated simulation analysis model of the wharf foundation and the actual engineering background, an overall multi-scale model of the wharf foundation is established, i.e., the final integrated simulation analysis model of the wharf foundation. Specifically, it includes: Step 1-2-2-1: Upper pier structure: construct the upper structure model and assign material properties, including density, Young's modulus, and Poisson's ratio of concrete; Step 1-2-2-2, pile foundation system: the upper wharf structure and the lower foundation are connected through the pile foundation system. The equivalent shear layer model is adopted, and the softening effect of the pile side friction resistance is simulated through zero-thickness cohesive elements. Step 1-2-2-3: For the lower soft soil foundation, assign soil parameters to foundation attributes, build a three-dimensional solid element model, and use a transition mesh in the area around the pile; Step 1-2-2-3, boundary condition setting, the upper pier structure adopts free field boundary, and the lower soft soil foundation applies viscoelastic boundary condition; Step 1-2-3, set the model loading strategy: A graded loading strategy was adopted to gradually apply the actual on-site load data to the integrated simulation analysis model of the wharf foundation to simulate the consolidation process and nonlinear response; The birth-death element method is used to pre-treat the foundation consolidation.

5. The method for integrated coupled modeling and simulation of a wharf and foundation considering large deformation and hysteresis characteristics of soil according to claim 4 is characterized in that: The construction of the multimodal data fusion network described in step 2 includes the following steps: Step 2-1, pre-processing the characteristic parameters output by the integrated simulation analysis model of the wharf foundation; Step 2-2: Build a multimodal data fusion network based on the Transformer network; Step 2-3: Dynamically calibrate and optimize the multimodal data fusion network.

6. The method for integrated coupled modeling and simulation of a wharf and foundation considering large deformation and hysteresis characteristics of soil according to claim 5 is characterized in that: The pre-processing of the characteristic parameters output by the integrated simulation analysis model of the wharf foundation described in step 2-1 includes: Wavelet transform is used to perform multi-scale decomposition processing on the characteristic parameters output by the integrated simulation analysis model of the wharf foundation and the real data obtained on site, as follows: Assume that the input data is x(t), and perform continuous wavelet transform on it as follows: Among them, a is the scale parameter, b is the translation parameter, ψ is the selected wavelet function, by selecting the scale range a∈[a min ,a max ] and reconstruction methods to filter out high-frequency noise and extract key feature parameters.

7. The method for integrated coupled modeling and simulation of a wharf and foundation considering large deformation and hysteresis characteristics of soil according to claim 6 is characterized in that: Step 2-2 builds a multimodal data fusion network, including the following steps: Step 2-2-1: Flatten the key feature parameter data of the spatial distribution to form a vector representation, specifically including: The embedding layer is used to encode the data sources of different key feature parameters to achieve unified dimensional mapping of the data; The self-attention module of the Transformer network is used to capture the correlation between time and space, compare the real-time monitoring data with key feature parameters, and extract key error information and structural response features; Step 2-2-2: Improve feature expression capabilities through residual connections and multi-head attention mechanisms; Step 2-2-3, optimize feature extraction and data enhancement, that is, introduce a feature extraction optimization module into the Transformer network to perform feature dimensionality reduction and information reconstruction on the fused multimodal data.

8. The method for integrated coupled modeling and simulation of a wharf and foundation considering large deformation and hysteresis characteristics of soil according to claim 7 is characterized in that: The dynamic correction and optimization of the multimodal data fusion network described in steps 2-3 includes the following steps: Step 2-3-1, prediction error acquisition and correction; Step 2-3-2, objective function and parameter tuning.

9. The method for integrated coupled modeling and simulation of a wharf and foundation considering large deformation and hysteresis characteristics of soil according to claim 8 is characterized in that: The prediction error acquisition and correction described in step 2-3-1 are as follows: The real-time monitoring data is compared with the characteristic parameters output by the integrated simulation analysis model of the wharf foundation to obtain the prediction error between the two. Assuming that the characteristic parameter output by the integrated simulation analysis model of the wharf foundation is y and the actual monitoring value is z, the state prediction process adopts the linear dynamic system model as follows: z m|m-1 =Ayz m-1|m-1 +Bl -1 P m|m-1 =AP k-1|k-1 From T +Q Among them, A is the state transfer matrix, B is the control matrix, l m-1 is the control input, Q is the process noise covariance matrix, P m∣m-1 is the prediction covariance matrix, z m∣m-1 is the prior prediction value of the monitoring value before time m, m represents the time of the mth prediction or observation, and k represents the time of the kth prediction or observation; The prediction error correction amount is calculated by the above-mentioned Kalman filter recursive formula, and the internal parameters of the model are dynamically adjusted so that the corrected prediction value is closest to the monitoring value.

10. The method for integrated coupled modeling and simulation of a wharf and foundation considering large deformation and hysteresis characteristics of soil according to claim 9, characterized in that: The objective function and parameter tuning described in step 2-3-2 specifically include: Construct the objective function and set the prediction output of the integrated simulation analysis model of the wharf foundation as f(x i ,θ), the actual output is o, then the objective function is defined as: By solving the above least squares problem, the parameter θ of the integrated simulation analysis model of the wharf foundation is updated to minimize the prediction error. Combined with the Bayesian updating theory, historical data and real-time data are used to continuously update the prior distribution and continuously correct the model parameter distribution.

Images