Intelligent construction simulation system for complex geological river-crossing tunnel deep foundation pit

Abstract

The application relates to the technical field of underground engineering and numerical simulation, and discloses a complex-geology-oriented intelligent construction simulation system for a river-crossing tunnel deep foundation pit, which comprises a data acquisition module, a data processing module, a topological evolution module, a parameter updating module, a solving engine, a boundary loading module and an internal memory interaction module; the data processing module generates an initial three-dimensional seepage hysteresis time matrix according to acquired tidal level and pore water pressure time series data; the topological evolution module calculates the shortest seepage distance of a grid based on an excavation face node; the parameter updating module combines the solving engine to output a soil volume strain calculation correction coefficient, generates an updated hysteresis time matrix and writes the hysteresis time matrix into an independent internal memory block; the solving engine combines the boundary loading module to read the hysteresis time amount, injects a transient boundary load and executes fluid-solid coupling calculation. The application converts a static boundary into a time hysteresis field which is dynamically updated according to construction excavation topologies and soil deformation, thereby improving the accuracy of fluid-solid coupling simulation calculation of a deep foundation pit under a tidal environment.

Description

Technical Field

[0001] This invention relates to the field of underground engineering and numerical simulation technology, specifically to an intelligent construction simulation system for deep foundation pits in cross-river tunnels with complex geological conditions. Background Technology

[0002] In the construction of deep foundation pit projects for cross-river tunnels, the foundation pits are often located in complex river geological environments. During the excavation, the tidal fluctuations of the external river surface will have a dynamic hydraulic effect on the foundation pit retaining structure and the surrounding soil. In order to assess the safety of the project construction process, fluid-structure interaction numerical simulation technology is used to analyze the stress and seepage state of the foundation pit.

[0003] However, existing numerical simulation methods have limitations in handling tidal water boundaries. Most existing models treat the external water body boundary as hydrostatic pressure that changes synchronously with the river surface water level or as a simple periodic dynamic load. This ignores the time lag effect caused by the propagation of tidal fluctuations in the soil pores. Even if some simulation systems consider the seepage lag time, they set static global boundary parameters at the beginning of the calculation. In the actual deep foundation pit excavation process, soil unloading will generate volumetric strain and cause changes in the void ratio. At the same time, the advancement of the excavation face will also directly change the length of the original seepage channel. This soil deformation and geometric boundary evolution caused by construction will change the propagation resistance and lag time of tidal fluctuations in the soil in real time. Traditional static boundary conditions cannot be adaptively updated with the construction steps, causing the pore water pressure field and effective stress field calculated by the model to gradually deviate from the actual engineering state.

[0004] Furthermore, during periods of dramatic tidal fluctuations, especially during low tide, traditional fluid-structure interaction boundary treatment methods can lead to unrealistic tensile stress calculations at nodes with excessively high local topography. This can affect the convergence and stability of the finite element matrix equations. Because existing simulation systems cannot accurately calculate the actual hydraulic gradient at each stage of excavation and track real-time changes in soil density, they struggle to accurately quantify the safety status of seepage failure risks such as piping and soil erosion, thus failing to provide reliable engineering early warnings for deep foundation pit construction. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides an intelligent construction simulation system for deep foundation pits in river-crossing tunnels with complex geological conditions. This system solves the problem in existing numerical simulations that cannot dynamically update the seepage conduction lag time based on changes in the foundation pit construction topology and soil deformation, leading to inaccurate fluid-solid coupling calculations under tidal boundaries.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an intelligent construction simulation system for deep foundation pits in river-crossing tunnels with complex geological conditions, the system being deployed in a computer device, the system comprising: The data acquisition module is used to synchronously acquire physical environment time-series data, which includes a continuous time series of absolute elevation of the river tide level and a time series of initial pore water pressure on the outside of the foundation pit retaining structure. The data processing module is used to extract the phase difference from the physical environment time series data through cross-correlation operation, and to perform a spatial three-dimensional interpolation algorithm to map it to the loaded grid nodes in the three-dimensional initial finite element mesh model to generate the initial three-dimensional penetration hysteresis time matrix. The topology evolution module is used to identify the grid nodes of the excavation face based on the construction step sequence, calculate the shortest effective seepage path distance from the retained loaded grid nodes to the grid nodes of the excavation face along the grid skeleton, and generate a grid distance matrix. The parameter update module is used to extract the volumetric strain increment of the soil, calculate the local phase offset correction coefficient based on the grid distance matrix and the volumetric strain increment to generate a diagonal correction matrix, multiply the initial three-dimensional permeability hysteresis time matrix with the diagonal correction matrix to generate an updated hysteresis time matrix, and write the updated hysteresis time matrix into an independent memory block. The solver engine is used to construct the three-dimensional initial finite element mesh model containing the foundation pit structure and soil-water boundary, and output the volumetric strain increment in the iterative calculation. The boundary loading module is used to read the lag time from the memory interaction module within the time integration iteration step, and inject load data into the global equilibrium equation system of the solver engine to perform fluid-structure interaction calculations based on the calculated transient boundary conditions. The memory interaction module is allocated with the aforementioned independent memory blocks for boundary data caching interaction.

[0007] By adopting the above technical solution, the initial three-dimensional seepage hysteresis time matrix is ​​constructed using a data processing module, the shortest seepage path changes caused by excavation are tracked using a topology evolution module, and the phase offset correction coefficient is calculated based on soil strain using a parameter update module. Therefore, the static seepage boundary is transformed into a time-varying boundary that is dynamically updated with grid topology reorganization and soil deformation, thus improving the boundary accuracy of deep foundation pit fluid-structure interaction simulation calculation.

[0008] Furthermore, the data acquisition module introduces a unified positioning system timestamp as a reference trigger signal to truncate and align the continuous time series of absolute elevation of the river surface tide with the initial pore water pressure time series on the time axis; the data acquisition module is equipped with a resampling unit, which performs cubic spline interpolation on the low-frequency data series to uniformly convert all input sequences into a set of equally spaced data points with the same discrete time step, generating a standardized digital matrix and outputting it to the data processing module.

[0009] By adopting the above technical solution, the time deviation caused by asynchronous triggering of different monitoring devices is eliminated, the sampling frequency of heterogeneous sensor data is unified, and a standardized data sequence is provided for performing frequency domain cross-correlation analysis and constructing a reference time field.

[0010] Furthermore, the data processing module performs a Fast Fourier Transform on the continuous time series of absolute elevation of the river tide and the initial pore water pressure time series, and filters out low-frequency trend terms and high-frequency environmental noise terms by setting a bandpass filter; calculates the cross-correlation function of the filtered tide series and each pore water pressure series in the time domain, and calculates the time offset corresponding to the peak value of the cross-correlation function as the phase difference of the main frequency band; the data processing module calls the three-dimensional ordinary kriging interpolation algorithm to calculate the interpolation weight coefficients of the loaded grid nodes for each discrete measuring point, and assembles the spatial interpolation results of the global nodes based on the interpolation weight coefficients to form the initial three-dimensional infiltration hysteresis time matrix.

[0011] By adopting the above technical solution, high and low frequency signal interference is filtered out in the frequency domain, the hysteresis time of tidal fluctuations in the soil pores is extracted, and the spatial interpolation algorithm is used to overcome the limitation of discrete sensor distribution, so as to realize the continuous mapping of time parameters in the three-dimensional coordinate system.

[0012] Furthermore, the topology evolution module extracts the set of geometric voxels of the soil mesh to be removed, performs a three-dimensional spatial Boolean difference operation in the finite element mesh coordinate system to extract the set of newly exposed excavation face mesh nodes; uses a fast traversal algorithm to solve the equation of process function, extracts the permeability coefficient of each continuous solid element, takes the reciprocal to generate a discrete slow-degree field, traces the shortest arrival path radiating from the loaded mesh node to the set of excavation face mesh nodes along the finite element mesh skeleton, obtains and assembles the shortest effective seepage path distance of each node to generate the mesh distance matrix of the current construction step.

[0013] By adopting the above technical solutions, the phenomenon of boundary retreat and seepage channel shortening caused by foundation pit excavation is quantified, and a spatial distance benchmark is established to assess the degree of influence of the internal unloading process on the seepage phase of the external boundary.

[0014] Furthermore, the parameter update module extrapolates the volumetric strain increment at the integration point to each loaded grid node, introducing an algebraic calculation based on the distance attenuation effect of the effective seepage path; the calculation process of the local phase shift correction coefficient includes: multiplying the volumetric strain increment by the strain time sensitivity coefficient, and multiplying by an exponential term, wherein the exponential term is based on the natural constant and has the negative of the ratio of the shortest effective seepage path distance to the characteristic influence length as the exponent, and subtracting the product obtained by the aforementioned multiplication from the value to obtain the local phase shift correction coefficient; and constructing the diagonal correction matrix using all the calculated local phase shift correction coefficients as the main diagonal elements.

[0015] By adopting the above technical solution, the volumetric deformation of the soil skeleton under fluid-structure interaction is converted into an algebraic parameter with a time delay dimension, reflecting the mechanical relationship between skeleton deformation and permeability change, and realizing nonlinear algebraic correction of seepage hysteresis field.

[0016] Furthermore, the parameter update module constructs a preset analytical function for future tidal fluctuations based on historical tidal hydrological data, and uses tidal harmonic analysis to decompose tidal fluctuations into a superposition of multiple simple harmonic waves with fixed periods. The parameter update module converts the updated lag time matrix and the feature parameter set of the preset analytical function for future tidal fluctuations into a continuous binary data stream, and directly writes the serialized binary data stream into the independent memory block in the physical running memory of the memory interaction module.

[0017] By adopting the above technical solutions, the input / output delay caused by the solver frequently reading external text files during time step iterations is avoided, and the computational coherence of the finite element system is improved by relying on memory mapping technology.

[0018] Furthermore, at the time integration point, the boundary loading module reads the feature parameter set and the updated lag time matrix from the independent memory block through the memory interaction module; extracts the seepage conduction delay time data corresponding to the loaded grid nodes; subtracts the current absolute integration time from the seepage conduction delay time data to obtain the effective tidal fluctuation time; and substitutes it into the preset future tidal fluctuation analytical function to solve for the equivalent absolute head elevation at the location of each loaded grid node.

[0019] By adopting the above technical solution, the global absolute time is converted into an independent effective response time based on the spatial location of the node, so that the hydraulic elevation applied to the equation system corresponds to the actual delay state of the model entity.

[0020] Furthermore, the boundary loading module introduces a truncation operator to limit the pore water pressure threshold. It calculates the difference between the absolute elevation of the equivalent head and the absolute vertical elevation of the loaded mesh node, compares this difference with the set minimum pore water pressure truncation threshold to select the maximum value, and then multiplies it by the volumetric unit weight of water to obtain the dynamic pore water pressure boundary value applied to the loaded mesh node. The dynamic pore water pressure boundary values ​​of all loaded mesh nodes are updated in the stiffness matrix and load vector of the finite element model to perform iterative solution.

[0021] By adopting the above technical solution, tensile stress that does not conform to physical reality is prevented when the external water surface is lower than the local node elevation during the low tide stage, thereby improving the iterative convergence stability of the fluid-structure interaction matrix equation under extreme water level conditions.

[0022] Furthermore, after completing the incremental iterative calculation of fluid-structure interaction, the solution engine performs a safety state determination, extracts pore water pressure data at each solid element to calculate the hydraulic gradient, extracts the volumetric strain data accumulated by the solid element in the current construction step, updates the current void ratio by combining it with the initial void ratio of the corresponding soil, and calculates the critical hydraulic gradient of the corresponding soil based on the relative density of the soil particles and the current void ratio.

[0023] By adopting the above technical solution, the soil compaction evolution caused by the unloading of the foundation pit excavation can be tracked, so that the judgment benchmark parameters for seepage failure can be matched with the soil state at the current construction stage in real time.

[0024] Furthermore, the solution engine calculates the ratio of the critical hydraulic gradient to the hydraulic gradient as the anti-seepage stability safety factor, and introduces a small protection constant to prevent overflow when dividing by zero in the hydraulic gradient calculation term in the denominator; the anti-seepage stability safety factor is numerically compared with a preset engineering safety factor threshold. When the anti-seepage stability safety factor of a local area is less than or equal to the engineering safety factor threshold, the corresponding entity unit is determined to be in a seepage failure risk state, and the spatial coordinates and physical quantity distribution matrix of the risk unit are extracted to output the rendered safety state cloud map.

[0025] By adopting the above technical solution, the calculation overflow caused by the hydraulic gradient approaching zero in the still water area inside the model is avoided, and the ability to resist seepage damage events such as piping and soil erosion in the whole domain is quantified.

[0026] This invention provides an intelligent construction simulation system for deep foundation pits in river-crossing tunnels with complex geological conditions. It offers the following advantages: 1. This invention extracts grid geometric reorganization information and soil volumetric strain data during the excavation process through a topology evolution module and a parameter update module, and updates the initial three-dimensional seepage lag time matrix into a dynamic lag time field. This processing method enables finite element simulation to synchronously reflect the actual impact of stratum excavation and deformation on seepage conduction path and time, overcomes the calculation errors caused by traditional static boundary conditions, and improves the boundary accuracy and overall precision of fluid-structure interaction calculation for deep foundation pits under tidal conditions.

[0027] 2. This invention utilizes a data processing module to perform frequency domain filtering and cross-correlation analysis on the collected time-series data. Combined with a spatial interpolation algorithm, it eliminates environmental noise interference and achieves continuous mapping of boundary parameters in three-dimensional space. At the same time, through a memory interaction module, it serializes and writes the analytical function parameters and the updated lag time matrix into physical memory, avoiding the delay caused by the solver frequently reading external files during the time integration iteration step, thereby improving the addressing rate of fluid-structure interaction boundary data and the continuity of the calculation process.

[0028] 3. This invention introduces a pore water pressure cutoff operator during the boundary loading stage to prevent unrealistic tensile stress disturbances caused by local node elevations exceeding the water surface during low tide, thereby enhancing the stability of the iterative solution of the matrix equation. In addition, the solution engine synchronously updates the current void ratio and critical hydraulic gradient of the soil based on the cumulative volumetric strain of the solid elements, and introduces a zero protection constant when calculating the anti-seepage stability safety factor. This enables the accurate quantification of the anti-seepage failure capability of the computational domain and the output of early warning information based on the actual soil properties of the current construction step. Attached Figure Description

[0029] Figure 1 This is a system architecture diagram of the present invention; Figure 2 This is a flowchart of the method of the present invention; Figure 3 This is a schematic diagram illustrating the principle of multi-source time-series data synchronization and baseline model construction in this invention. Figure 4 This is a schematic diagram illustrating the principle of frequency domain cross-correlation analysis and lag time field construction of the present invention. Figure 5 This is a schematic diagram illustrating the excavation topology evolution and seepage path addressing principle of the present invention; Figure 6 This is a schematic diagram illustrating the derivation principle of the phase offset correction coefficient under stress-seepage coupling in this invention. Figure 7 This is a schematic diagram illustrating the principle of adaptive update of the lag time matrix and memory serialization in this invention. Figure 8 This is a schematic diagram of the dynamic boundary condition reconstruction and fluid-structure interaction solution principle of the present invention; Figure 9This is a schematic diagram illustrating the engineering safety status assessment and seepage failure early warning principle of the present invention. Figure 10 This is a schematic diagram of the planar topology of the deep foundation pit monitoring points for the river-crossing tunnel according to the present invention; Figure 11 This is a time-history comparison curve of pore water pressure at the observation point as a function of tidal fluctuations, as presented in this invention. Figure 12 This is a scatter plot comparing the spatial distribution of surface settlement around the foundation pit of this invention.

[0030] The module consists of: 10. Data acquisition module; 20. Data processing module; 30. Topology evolution module; 40. Parameter update module; 50. Solver engine; 60. Boundary loading module; and 70. Memory interaction module. Detailed Implementation

[0031] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0032] See attached document Figure 1 This invention provides an intelligent construction simulation system for deep foundation pits in cross-river tunnels with complex geological conditions. The system includes: a data acquisition module 10, a data processing module 20, a topology evolution module 30, a parameter update module 40, a solution engine 50, a boundary loading module 60, and a memory interaction module 70.

[0033] The data acquisition module 10 is deployed at the construction site and surrounding water area to synchronously acquire continuous physical environment time-series data. The data processing module 20 is connected to the data acquisition module 10 to receive the time-series data and perform frequency domain cross-correlation calculations and spatial field interpolation.

[0034] The topology evolution module 30 is configured in the preprocessing layer of the solver engine 50 and is used to perform three-dimensional mesh stripping and seepage path addressing based on the construction steps. The parameter update module 40 connects the topology evolution module 30 and the data processing module 20 and extracts the calculation results output by the solver engine 50 for tensor algebra operations.

[0035] The solver engine 50 includes a boundary loading module 60. The parameter update module 40 serializes and stores the calculated boundary control matrix through the memory interaction module 70. During the integral step of the solver, the boundary loading module 60 initiates a low-level addressing to the memory interaction module 70 and injects the resolved transient load into the global stiffness and permeability matrix system of the solver engine 50.

[0036] See attached document Figure 2 This invention provides an intelligent construction simulation method for deep foundation pits in river-crossing tunnels with complex geological conditions, comprising the following steps: S100 and data acquisition module 10 simultaneously acquire the continuous time series of absolute elevation of the river tide level and the initial pore water pressure time series at different spatial locations outside the foundation pit retaining structure. Solver 50 constructs a three-dimensional initial finite element mesh model including the foundation pit retaining structure, internal soil, surrounding soil and river water boundary. S200 and data processing module 20 perform Fourier transform and cross-correlation calculations on the collected tidal level and pore water pressure time series data to extract the phase difference, and execute a spatial three-dimensional interpolation algorithm to map the discrete phase difference to the finite element loaded mesh nodes to generate the initial three-dimensional infiltration hysteresis time matrix. S300. In the pre-processing of construction steps, the topology evolution module 30 extracts the geometric voxels of the soil mesh to be removed according to the construction step sequence and performs spatial Boolean difference set operation to identify the mesh nodes of the excavation face. The shortest effective seepage path distance of the retained load-bearing nodes along the mesh skeleton is calculated with the mesh nodes of the excavation face as the source point, and a mesh distance matrix is ​​generated. S400 and parameter update module 40 extract the volumetric strain increment of the global grid Gaussian points after the convergence of the fluid-structure interaction calculation in the previous construction step, calculate the local phase offset correction coefficient based on the grid distance matrix and the volumetric strain increment, and generate the diagonal correction matrix. S500, the parameter update module 40 performs matrix multiplication on the initial three-dimensional permeation lag time matrix and the diagonal correction matrix to generate an updated lag time matrix, and writes the updated lag time matrix and the preset future tidal fluctuation analysis function into the independent memory block of the memory interaction module 70. S600, the solver engine 50 starts the fluid-structure interaction sequential solution. In the time integration iteration step, the boundary loading module 60 initiates direct addressing to read the lag time from the memory interaction module 70. Based on the time domain translation principle, the transient head elevation value is calculated as a forced head vector element and injected into the global equilibrium equation system. S700 and the solver engine 50 calculate and converge, and output nodal displacement, pore pressure distribution and volumetric strain tensor. The system decides to enter the next excavation construction step cycle or terminate the simulation process based on the working condition judgment conditions.

[0037] See attached document Figure 3 Before performing finite element calculations, it is necessary to establish the underlying mesh structure and standardize the external environmental input sources.

[0038] Based on the engineering geological survey report and main structure design drawings, the solver engine 50 constructs a three-dimensional initial finite element mesh model that includes the foundation pit retaining structure, internal soil, surrounding soil, and river water boundary. During the mesh discretization process, the soil region adopts fluid-structure interaction solid elements with dual degrees of freedom of displacement and pore water pressure to support the stress-seepage synchronous solution in subsequent calculations. The foundation pit retaining structure is spatially modeled using solid elements or shell elements. The global coordinate system of the finite element mesh model is established based on the absolute coordinate origin specified in the construction site.

[0039] For the initial geostress equilibrium process of the finite element model, the self-weight stress field and the static pore water pressure field determined by the initial groundwater level can be set according to the actual engineering soil unit weight and the preset void ratio, and the displacement zeroing operation can be performed to establish a real initial fluid-structure interaction equilibrium state.

[0040] Data acquisition module 10 deploys a physical monitoring network at the construction site and surrounding waters. It also installs a tide gauge on a fixed structure on the river surface near the foundation pit, continuously recording the absolute elevation changes of the river water level to form a continuous time series of absolute tide levels, defined as... .

[0041] Among them, the environmental parameters have the characteristic of fluctuating with time. The data acquisition of the tide gauge reflects the periodic changes in the river water level. The data acquisition module 10 drills holes in the soil on the outer perimeter of the water-facing side of the foundation pit retaining structure and buries multiple pore water pressure gauges at different depths and horizontal orientations.

[0042] To capture the three-dimensional spatial variability of the underground seepage field, the horizontal spacing between adjacent pore water pressure gauges was set to 10 to 50 meters, with the burial depth covering a range from 5 meters below the surface to 20 meters below the maximum excavation depth of the foundation pit. Each pore water pressure gauge continuously recorded the pore pressure fluctuations within the surrounding soil, forming an initial pore water pressure time series, defined as... Each deployed pore water pressure gauge records three-dimensional spatial coordinates. The physical coordinates are converted to coordinate points corresponding to the three-dimensional initial finite element mesh model constructed by the solver engine 50.

[0043] In the above time series data definition: The absolute time variable of physical monitoring; The spatial index number of the pore water pressure gauge, with a value of =1,2,…, ; The total number of pore water pressure gauges deployed is determined based on the size of the foundation pit; for example, the value range can be from 5 to 50. The first The horizontal, vertical, and depth coordinates of a pore water pressure gauge in the global coordinate system.

[0044] The raw time-series data acquired by the sensing hardware often has time axis offset due to different device triggering mechanisms during actual operation. The data acquisition module 10 introduces a unified global positioning system timestamp as a reference trigger signal to truncate and align the data streams recorded by the tide gauge and pore water pressure gauge on the time axis, and extract the effective data segments within the same continuous natural time period.

[0045] Because the default sampling frequencies of the sensors for river surface tide level and underground pore water pressure are different, a resampling unit is configured in the data acquisition module 10 to ensure the accuracy of subsequent frequency domain analysis. The resampling unit performs cubic spline interpolation on the lower-frequency data sequence. Cubic spline interpolation constructs piecewise cubic polynomials between adjacent data points, ensuring the fitted curve remains continuous at nodes with continuous first and second derivatives, thus smoothly filling in missing data values. This is particularly relevant for adjacent time nodes in a given low-frequency data sequence. and The interpolation polynomial function constructed by the resampling unit is expressed as: ; In the formula, For a moment The corresponding resampling interpolation value; These are known adjacent sampling time nodes in the low-frequency sequence; For physical monitoring, the absolute time variable is... ; , , , These are the piecewise polynomial coefficients obtained based on the function values ​​of adjacent nodes and the boundary conditions of derivative continuity.

[0046] Through the above calculations, the resampling unit uniformly converts all input sequences into sequences with the same discrete time step. A set of equally spaced data points, parameters The fixed time interval after resampling is set to a range of 1 to 10 minutes to cover the main frequency band of tidal fluctuations. After time synchronization and resampling, the continuous time series of absolute elevation of the river surface tide and the initial pore water pressure time series are formatted into equal-length digital matrices aligned by the time dimension. The standardized digital matrix is ​​output by the data acquisition module 10, which constitutes the baseline data source for subsequent extraction of physical field hysteresis features.

[0047] See attached document Figure 4In actual geotechnical engineering, the soil has a natural pore damping effect on the seepage conduction of fluids. This damping is macroscopically manifested as the time lag when external water level fluctuations are conducted to underground structures. In order to convert the physical delay obtained in engineering environmental monitoring into computational parameters that can be read by the finite element bottom mesh, the data processing module 20 receives the standardized digital matrix and extracts the hysteresis characteristics of tidal fluctuations conducted inside the soil.

[0048] In a preferred embodiment, the data processing module 20 performs a Fast Fourier Transform on the continuous time series of absolute tidal elevation and the initial pore water pressure time series. The data processing module 20 filters out low-frequency trend terms and high-frequency environmental noise terms using a bandpass filter. During this processing, the data processing module 20 sets the lower limit of the bandpass filter to 1.0 × 10⁻⁶ based on the local diurnal or full-diurnal tidal cycle characteristics. -5 Hertz, with an upper limit set at 3.0 × 10⁻⁶. -5 Hertz is used to obtain the main frequency band data that reflects the real physical seepage. The data processing module 20 calculates the cross-correlation function between the filtered tidal level sequence and each pore pressure sequence in the time domain. For the discrete time series obtained by the digital sampling device, this cross-correlation function is equivalent to the multiplication and accumulation of the data corresponding to the time point in the engineering calculation platform. The integral calculation formula is expressed as: ; In the formula, For the first The time-domain cross-correlation function values ​​corresponding to each pore water pressure gauge; This is a time offset variable, with a value range set from -12 hours to 12 hours based on the semi-diurnal tide cycle span; This is a continuous time series of absolute elevations of the river surface tide after alignment and filtering. For the aligned and filtered first Time series of initial pore water pressure at each location; For physical monitoring, the absolute time variable; Spatial index number for the pore water pressure gauge; The total time window length for truncating and aligning the extracted effective data segments is set to a range of 72 to 360 hours to ensure stable extraction of tidal cycle characteristics.

[0049] The cross-correlation function value reflects the waveform similarity of two time series at different time offsets. The data processing module 20 iterates through the time offset variables. The value range of is used to address and calculate the time offset corresponding to the peak value of the cross-correlation function. This time offset represents the transmission of tidal head load through formation pores to the first tidal head load. The data processing module 20 defines the extracted time delay at the location of each pore water pressure gauge as the main frequency band phase difference, and the extraction formula is as follows: ; In the formula, For the transmission of tidal fluctuations to the first The phase difference of the corresponding main frequency band of each sensor; A mathematical operator for the independent variable that yields the maximum value of a function; For the first The time-domain cross-correlation function values ​​corresponding to each pore water pressure gauge.

[0050] Due to limitations in project cost and construction conditions, the pore water pressure gauges are distributed as discrete points around the foundation pit, which cannot directly cover all loaded boundary nodes in the finite element model generated by the solver engine 50. The data processing module 20 calls the three-dimensional ordinary kriging interpolation algorithm to map the discrete main frequency band phase difference to the entire model domain. The three-dimensional ordinary kriging interpolation algorithm constructs a variogram model in space to quantify the evolution of the correlation of the phase difference in three-dimensional space as the physical distance decreases. For the fitting of the variogram model and the derivation and solution of the kriging equations, a spherical model or an exponential model can be selected for calculation based on the spatial distribution characteristics. When fitting the variogram model, based on the spatial scale characteristics of the foundation pit area, the range of the variogram parameter is set to 50.0 meters to 500.0 meters to reasonably characterize the spatial autocorrelation distance of the seepage phase.

[0051] Data processing module 20 calculates the interpolation weighting coefficients of the target loaded grid nodes for each discrete measuring point based on the three-dimensional spatial coordinates of each pore water pressure gauge and the corresponding dominant frequency band phase difference. To ensure the unbiasedness of spatial interpolation, the calculated interpolation weighting coefficients must satisfy the algebraic constraint that the sum equals 1. Data processing module 20 generates the spatial interpolation results for the entire domain nodes based on the calculated weighting coefficients. The calculation formula is expressed as follows: ; In the formula, The interpolation calculation is then mapped to the finite element loaded mesh nodes. The initial time delay value; To determine the index number of the loaded mesh nodes in engine 50; This represents the total number of pore water pressure gauges deployed. For the first A pore water pressure gauge for the target loaded grid node Kriging space interpolation weighting coefficients; For the transmission of tidal fluctuations to the first The phase difference of the corresponding main frequency band of each sensor; This is the spatial index number for the pore water pressure gauge.

[0052] Data processing module 20 vectorizes and assembles the calculated initial time delay values ​​according to the numbering order of the loaded grid nodes. The assembled data structure constitutes the initial three-dimensional seepage hysteresis time matrix. This matrix digitally represents the spatial distribution characteristics of the time hysteresis of the seepage transmission of tidal water flow load to the surrounding strata in the initial state before excavation of the foundation pit, and is defined as follows: The dimensions of the matrix are set to... ,in This represents the total number of loaded mesh nodes in the finite element model, with values ​​ranging from 10. 3 Up to 10 6 The specific values ​​depend on the discrete density of the grid and the model scale. The initial three-dimensional permeation hysteresis time matrix, as the physical reference field data, is transmitted and temporarily stored in memory to support the dynamic evolution calculation of multi-physics coupling iteration under subsequent excavation conditions.

[0053] See attached document Figure 5 During the continuous excavation of the foundation pit under multiple working conditions, the continuous removal of underground soil will change the original geometric boundaries and seepage paths of the strata. In order to update this part of the geometric topology and re-evaluate the seepage conduction path in the calculation model, the topology evolution module 30 enters the... During the pre-processing stage of the construction step, the geometric topology of the finite element mesh is deduced, in which... It is a positive integer greater than or equal to 1.

[0054] The topology evolution module 30 extracts the first step according to the preset construction sequence. In the construction step, the set of geometric voxels of the soil mesh to be removed is used, and the topology evolution module 30 performs three-dimensional spatial Boolean difference operations in the finite element mesh coordinate system.

[0055] The topology evolution module 30 calculates the difference between the initial global mesh set or the mesh set retained from the previous working condition and the current soil mesh set to be removed. By changing the mesh material properties or using the birth and death element technique, the elements to be excavated are extracted from the topology. After Boolean difference operation, some mesh nodes that were originally inside the soil will be exposed to the free surface. The topology evolution module 30 identifies and extracts the set of newly exposed excavation face mesh nodes in the current step, defining them as... This set constitutes the physical boundary for the current groundwater seepage accumulation during the construction phase.

[0056] Because the unloading and removal of the soil inside the excavation pit will cause stress redistribution in the surrounding retained soil and change the dissipation path length of pore water pressure along the soil skeleton, in order to quantify the spatial impact of boundary changes on fluid conduction distance, the topology evolution module 30 uses the excavation face grid node set. Given the set of source points, the computational model retains the shortest topological distance along the effective seepage path for all loaded nodes within the domain.

[0057] The topology evolution module 30 employs a fast traversal algorithm to solve the equations. This algorithm, by simulating wavefront expansion, traces the shortest path of radiation outward from the source point set along the finite element mesh skeleton. To ensure that the partial differential equations have unique solutions and to initiate traversal addressing, the topology evolution module 30 sets initial boundary conditions at the source points. The differential expression and boundary conditions of the equations in three-dimensional space are as follows: ; ; In the formula, 3D coordinate nodes To the excavation face grid node set The shortest equivalent distance for seepage; These are the spatial coordinates of the finite element mesh nodes in the global coordinate system. The mathematical notation for calculating the vector norm, representing the magnitude of the gradient; It is a three-dimensional spatial gradient operator used to calculate the rate of change of the distance field in various spatial directions; The spatial slowness function of the medium; This refers to the set of newly exposed excavation face mesh nodes in the current step.

[0058] In fluid-structure interaction analysis scenarios, the spatial slowness function Characterizing the time cost or resistance required for a fluid to travel a unit distance through a medium, its value is determined based on the reciprocal of the permeability coefficient matrix of anisotropic soil layers. The parameter varies depending on the type of geological structure. The value range is set to 1.0 × 10. -2 Up to 1.0×10 6 Between seconds and meters.

[0059] At the mesh level, the topology evolution module 30 extracts the permeability coefficient of each continuous solid element, takes the reciprocal and discretizes it to the corresponding mesh node to form a discrete slow field. The specific windward differential discretization format and narrowband extension update strategy of the fast travel algorithm can be implemented by conventional programming based on the finite element mesh structure.

[0060] By solving the above functional equation, the topology evolution module 30 obtains the globally retained loaded nodes. Along the continuous mesh skeleton to the set of mesh nodes at the excavation face The shortest effective seepage path distance is defined as The topology evolution module 30 performs vector assembly of the shortest effective seepage path distances for each loaded node according to the grid node numbering order, generating the current topology evolution module. The grid distance matrix for each construction step is defined as follows: .

[0061] The data dimensions of the grid distance matrix are consistent with those of the initial three-dimensional penetration hysteresis time matrix constructed in step two, and are both configured as follows: ,in This represents the total number of loaded mesh nodes in the finite element model. Since the excavation operation only removes the internal soil mesh, the total number of loaded nodes at the outer boundary is [not specified]. The distance matrix remains unchanged during the construction process, which ensures the correspondence of dimensions during subsequent matrix algebra operations. The assembled grid distance matrix is ​​output and transmitted to the parameter update module 40 as the underlying spatial mapping parameter to support subsequent time offset correction calculations based on the soil deformation field.

[0062] See attached document Figure 6 Since deep foundation pit engineering involves complex fluid-structure interaction, soil excavation will lead to the release of stratum stress, which in turn will cause volumetric deformation of the soil skeleton. This volumetric strain is physically manifested as the increase or decrease of soil void ratio and the deformation or closure of local pore channels. Changes in pore structure will directly lead to changes in permeability, causing the actual time delay of tidal water flow in the soil to shift dynamically relative to the initial state before excavation. In order to effectively correct the nonlinear error caused by this physical process in numerical simulation, the parameter update module 40 converts mechanical deformation into correction parameters for boundary time offset.

[0063] The parameter update module 40 retrieves the previous construction step, i.e., the first step, from the calculation result database of the solver engine 50. The construction step involves the global soil volumetric strain increment data after the fluid-structure interaction calculation has converged. It should be noted that this data is generated when the simulation is in the first excavation construction step. At that time, since no excavation disturbance has occurred, the system sets the initial value of the volumetric strain increment of the entire soil to zero. In the finite element analysis mechanism, the derived physical quantities such as strain and stress are solved and stored in the internal Gaussian integration points of each element during the solution process.

[0064] To align the deformation data within the element with the data dimensions of the boundary nodes, the parameter update module 40 performs a data mapping operation, extrapolating the volumetric strain increment at the Gaussian integration point to each loaded mesh node. The specific algorithm implementation for the extrapolation of finite element shape functions and the smoothing of nodal physical quantities can be configured based on conventional finite element post-processing theory.

[0065] After data smoothing, parameter update module 40 obtains the global load nodes. The incremental volumetric strain at each node generated in the previous construction step, combined with the current working condition grid distance matrix output by the topology evolution module 30, is used by the parameter update module 40 to calculate the local phase shift correction coefficient for each loaded node. This calculation logic introduces a distance attenuation effect based on the effective seepage path, so that areas closer to the excavation face and with more obvious deformation receive corresponding correction weights. The formula for calculating the local phase shift correction coefficient is as follows: ; In the formula, For the calculation of the obtained loaded mesh nodes The local phase shift correction coefficient is set with a minimum cutoff threshold of 0.1 to prevent non-physical negative time delays caused by extreme soil expansion. The strain-time sensitivity coefficient characterizes the proportional relationship between the change in seepage lag time caused by the increase in strain per unit volume. It is calibrated based on the seepage consolidation test data during the engineering geological exploration stage, and its value ranges from 10.0 to 100.0. For loaded mesh nodes The extrapolated volumetric strain increment is positive for tensile expansion and negative for compressive consolidation. For the natural constant Exponential function operators with base 0; Loaded grid nodes extracted from the grid distance matrix The shortest effective seepage path distance; The characteristic influence length characterizes the effective range of the impact of excavation disturbance on the seepage path. It is set according to the excavation depth of the foundation pit and ranges from 10.0 meters to 50.0 meters. To solve for the index number of the loaded mesh node in engine 50.

[0066] The parameter update module 40 traverses all loaded mesh nodes on the outer boundary of the finite element model and calculates their respective local phase offset correction coefficients one by one. The parameter update module 40 arranges all the obtained local phase offset correction coefficients in the order of the mesh node numbers and constructs a square matrix as the main diagonal elements.

[0067] This square matrix is ​​defined as a diagonal correction matrix, denoted as: The dimensions of the diagonal correction matrix are configured as follows: ,in The total number of loaded mesh nodes in the finite element model is represented by the matrix. All off-diagonal elements in the matrix are set to 0. The diagonal correction matrix vectorizes the time offset caused by stress-seepage coupling and uses it as the core algebraic parameter for subsequent adaptive updates of the seepage time boundary.

[0068] See attached document Figure 7After obtaining the diagonal correction matrix that reflects soil deformation, the parameter update module 40 performs algebraic operations to adaptively update the seepage time delay parameters of each loaded grid node, and sends the updated data to the underlying memory structure.

[0069] At the physical level, the reorganization or volume expansion of the soil's internal pore structure caused by the excavation of the foundation pit will directly change the permeation resistance of the fluid in the porous medium. In order to quantify the impact of this deformation on time lag in the numerical model, the parameter update module 40 calls the initial three-dimensional permeation lag time matrix and performs matrix multiplication with the calculated diagonal correction matrix.

[0070] Algebraically, since all off-diagonal elements of the diagonal correction matrix are zero, this matrix multiplication is essentially equivalent to independently scaling the values ​​of each node in the initial time delay field locally linearly. Through this operation, the initial reference lag time field is compensated for the dynamic offset caused by the excavation deformation of the foundation pit. The formula for the matrix multiplication operation is as follows: ; In the formula, The updated lag time matrix has the following dimension configuration: The diagonal correction matrix generated by extracting global strain data for parameter update module 40 is configured with the following dimensions. The initial three-dimensional seepage hysteresis time matrix contains the spatial distribution characteristics of time hysteresis in the initial state before excavation of the foundation pit, with the dimensions configured as follows: The total number of externally loaded grid nodes in the model is represented by the updated hysteresis time matrix generated by the above calculations, which encapsulates the corrected seepage conduction delay time data of all loaded nodes in the current construction step.

[0071] In finite element transient analysis, the head load at the outer boundary changes continuously in the time integration domain. To provide dynamic boundary conditions for subsequent calculation steps, parameter update module 40 constructs a preset analytical function for future tidal fluctuations based on historical tidal hydrological data. This analytical function is established using tidal harmonic analysis, decomposing complex tidal level fluctuations into a superposition of multiple simple harmonic waves with fixed periods, thereby achieving function extrapolation on the time axis. The mathematical expression of the preset analytical function for future tidal fluctuations is: ; In the formula, absolute time The absolute elevation of the river tide level as predicted at all times; The absolute elevation of the local mean sea level is calculated from the arithmetic mean of the continuous time series of absolute elevations of the river tide level extracted by the data acquisition module 10 or the zero-frequency DC component of the fast Fourier transform. The total number of selected tidal harmonic components is set to a range of 4 to 11 to ensure fitting accuracy and computational efficiency. Index number for the tidal harmonic components; For the first The amplitude of the harmonic component, For the first The angular frequency of the harmonic component; For the first The initial phase of each harmonic component; For physical monitoring, the absolute time variable is denoted by , and for the integration process of the matching solver, the value range is set to the current . The physical time span of a construction step is from the start time to the planned completion time of that construction step. The physical time span of a single construction step is set according to the actual amount of excavation work, and the value ranges from 12 hours to 168 hours.

[0072] The least squares fitting process for the amplitude, angular frequency and initial phase of each harmonic component can be programmed based on conventional tidal analysis theory.

[0073] Considering that conventional finite element engineering software frequently reads static text files from external hard drives to update boundary conditions in small-step iterative solutions of multiphysics coupling, which can lead to communication delays in the computer's low-level input and output, in a preferred embodiment of the present invention, the parameter update module 40 communicates with the memory interaction module 70 through direct memory mapping technology.

[0074] The parameter update module 40 converts the updated lag time matrix and the characteristic parameter set of the preset future tidal fluctuation analytical function into a continuous binary data stream and performs a data serialization operation. The system allocates an independent memory block in the physical running memory for the memory interaction module 70. The parameter update module 40 calls the operating system's low-level application programming interface to directly write the serialized binary data stream to a specified memory address in the independent memory block. This data writing mechanism refreshes the external load data source for the finite element solution, ensuring that the boundary subroutine embedded in the subsequent solution engine 50 can be quickly addressed and read in the integration iteration step, thereby improving the overall efficiency of model calculation.

[0075] See attached document Figure 8 After completing the data serialization in the underlying memory, the finite element model is ready to enter the nonlinear solution stage of the current construction step. In order to reflect the lag effect of the tide in space in the time integration domain, the boundary loading module 60 dynamically reconstructs the boundary conditions of the loaded nodes based on the data in memory and drives the solution engine 50 to complete the fluid-structure interaction calculation.

[0076] At each discrete-time integration point in the finite element transient analysis, the solution engine 50 triggers the embedded boundary loading module 60. The boundary loading module 60 reads data from an independent memory block through the memory interaction module 70 to obtain the parameter set of the preset future tidal fluctuation analytical function and the updated lag time matrix.

[0077] For any loaded grid node on the outer boundary The boundary loading module 60 extracts the corresponding node seepage conduction delay time data from the updated hysteresis time matrix. Based on the current absolute integral time, the boundary loading module 60 calculates the effective tidal fluctuation time experienced by the node, expressed by the following formula: ; In the formula, For loaded mesh nodes Effective tidal fluctuation time; To solve for the absolute integral time of engine 50 in the current increment step; This is the node seepage conduction delay time data extracted from the updated hysteresis matrix.

[0078] The boundary loading module 60 substitutes the calculated effective tidal fluctuation time into the preset future tidal fluctuation analytical function to solve for the equivalent external head elevation at the corresponding node position at that moment. The calculation formula is expressed as follows: ; In the formula, To calculate the loaded mesh nodes The equivalent absolute elevation of the head at the current absolute integration time; The absolute elevation of the local mean sea level benchmark; The total number of tidal harmonic components; Index number for the tidal harmonic components; For the first The amplitude of each harmonic component; For the first The angular frequency of the harmonic component; For the first The initial phase of each harmonic component; For loaded mesh nodes Effective tidal fluctuation time.

[0079] After obtaining the equivalent absolute head elevation, the boundary loading module 60 converts it into dynamic pore water pressure applied to the nodes of the finite element loaded mesh based on the principle of hydrostatics. During the receding tide, the water level will drop below the node elevation, at which point the surface of the node area is exposed and unpressurized. To avoid introducing unrealistic negative tensile stress into the finite element stiffness equation and causing calculation divergence, the boundary loading module 60 introduces a truncation operator for threshold limitation. The conversion formula for pore water pressure is expressed as: ; In the formula, To apply to the loaded mesh nodes Dynamic pore water pressure boundary value; The volumetric density of water is determined based on the physical properties of local groundwater and river water, and its value ranges from 9.8 kN / m³ to 10.1 kN / m³. This is a maximum value function operator used to select the term with the largest algebraic value from multiple variables; This is the absolute elevation of the equivalent head; For loaded mesh nodes Vertical absolute elevation in the global coordinate system; 0 is the minimum cutoff threshold for pore water pressure, used to prevent physical tension when the water surface is below the node elevation.

[0080] The boundary loading module 60 traverses all loaded mesh nodes and updates the calculated dynamic pore water pressure boundary values ​​to the stiffness matrix and load vector of the finite element model. Based on this, the solution engine 50 performs the Newton-Raphson incremental iterative solution of the fluid-structure interaction equation of porous media. During the nonlinear iteration process, the system determines the numerical convergence state by calculating the global displacement increment norm and the unbalanced force norm.

[0081] To balance solution accuracy and computation time, the preset tolerance standard configuration in this embodiment is: the relative displacement tolerance range is set at 1.0 × 10⁻⁶. -4 Up to 5.0×10 -3 The relative tolerance range for unbalanced forces is set at 1.0 × 10⁻⁶. -3 Up to 1.0×10 -2 Between these points, when the global displacement increment and the unbalanced force meet the above-mentioned preset tolerance criteria, the solver engine 50 determines that the calculation of the current time step has converged and outputs the displacement, stress and pore water pressure field data of the global mesh.

[0082] After completing the calculation for the current time step, the solver engine 50 determines the current time step. The system checks whether the simulation time for the construction step has been completed. If it has, and there are subsequent construction plans, the system proceeds to the next step. The system performs construction steps and then calls the topology evolution module 30 to start the next round of mesh boundary subdivision and seepage path update, thus forming a cyclical iteration of construction evolution. If all construction steps of the foundation pit are simulated and completed, the system stops incremental calculation, archives the evolution data of the entire working condition, and outputs the final engineering evaluation report.

[0083] See attached document Figure 9 After the solver engine 50 completes the incremental iterative calculation of fluid-structure interaction for all working conditions, the system obtains a complete evolution dataset containing the mesh displacement, stress and pore water pressure in the entire spatial domain.

[0084] To guide the construction sequence of actual foundation pit projects and provide auxiliary decision-making, the solver engine 50 performs an engineering-level safety status assessment on the calculation results when entering the post-processing stage, and outputs warning information when the assessment results exceed the safety tolerance range.

[0085] In deep foundation pit projects near water bodies, fluctuations in external tidal water levels create a head difference between the inner and outer sides of the foundation pit's water-retaining structure. This head difference drives groundwater seepage, which in turn generates seepage dynamics within the underground soil. When the magnitude of the seepage dynamics exceeds the effective unit weight of the soil skeleton, it can trigger seepage failure phenomena such as piping and soil erosion.

[0086] To this end, the solver engine 50 extracts pore water pressure data at the integration points of each solid element in the finite element model, calculates the spatial pore water pressure gradient, and converts it into a hydraulic gradient. The formula for calculating the hydraulic gradient is as follows: ; In the formula, For the first Hydraulic gradient of each solid unit; For the first The spatial pore water pressure gradient vector of each solid element is obtained by differentiating the nodal pore pressure values ​​at each vertex of the element based on the geometric shape function. A mathematical operator for calculating the magnitude of a vector; This is the volumetric specific gravity of water, ranging from 9.8 kN / m³ to 10.1 kN / m³. This is the index number of the entity unit.

[0087] Based on the physical parameters obtained during the geological exploration phase, the solver engine 50 calculates the critical hydraulic gradient of local soil to resist seepage failure. The critical hydraulic gradient is controlled by the physical compaction state of the soil itself, and the calculation formula is expressed as follows: ; In the formula, For the first The critical hydraulic gradient of the soil corresponding to each solid element; For the first The relative density of soil particles at each physical unit location ranges from 2.6 to 2.8. This corresponds to the current void ratio of the soil after volumetric deformation at the current construction step.

[0088] To accurately obtain the current void ratio, the solution engine 50 extracts the current construction step. The cumulative volumetric strain data of each solid element is combined with the initial void ratio of the soil for algebraic evolution update. The update formula for the current void ratio is expressed as: ; In the formula, For the first Each solid element corresponds to the current void ratio of the soil at the current construction step; For the first Each solid unit corresponds to the initial void ratio of the soil before excavation disturbance, which is obtained based on undisturbed geotechnical tests and ranges from 0.4 to 1.5. For the first The cumulative volumetric strain of a solid element from its initial state to the current construction step is positive for tension and negative for compression. Furthermore, the calculated current void ratio... If the soil undergoes severe compression deformation that is less than the theoretical lower limit (e.g., 0.1) or excessive stretching expansion that is greater than the theoretical upper limit (e.g., 3.0), the system will forcibly assign a lower limit of 0.1 or an upper limit of 3.0 to maintain the stability of subsequent effective unit weight and hydraulic gradient calculations and prevent non-physical calculation divergences caused by extreme deformation.

[0089] After obtaining the actual hydraulic gradient and critical hydraulic gradient of the unit, the solution engine 50 calculates the anti-seepage stability safety factor to quantify the safety margin. To avoid the division-by-zero overflow anomaly caused by the hydraulic gradient approaching zero under still water conditions, and to ensure the logical completeness of the algorithm under various hydraulic boundaries, the solution engine 50 introduces a small protection constant in the denominator term. The calculation formula is expressed as: ; In the formula, For the first The anti-permeability stability safety factor of each physical unit; For the first The critical hydraulic gradient of the soil corresponding to each solid element; For the first Hydraulic gradient of each solid unit; To prevent the denominator from overflowing when divided by zero, the value is set to 1.0 × 10⁻⁶. -6 .

[0090] The solution engine iterates through all entity cells within the computational domain 50 times, calculating the anti-permeability stability safety factor for each cell. Compared with the preset engineering safety factor threshold Numerical comparisons were performed, and the preset engineering safety factor threshold was determined. The value range is set from 1.5 to 2.0, and the specific value is configured according to the relevant engineering specifications and the safety level standards of the foundation pit.

[0091] When the anti-permeability stability safety factor of a certain unit Less than or equal to the preset engineering safety factor threshold When the problem is solved, the solver engine 50 determines that the local area is in a state of potential seepage failure risk and temporarily records the spatial coordinates of the unit, the current construction step number and the risk data.

[0092] The post-processing interface of the system's underlying program extracts the risk unit coordinate set and physical quantity distribution matrix output by the solver engine 50. The system transmits the relevant data to an external display terminal and uses conventional graphics rendering technology to generate a three-dimensional safety status cloud map containing displacement contour lines and seepage risk highlight marks. The color mapping matrix generation and graphics rendering acceleration algorithm for the three-dimensional cloud map can be developed based on conventional computer graphics post-processing frameworks. The system finally outputs the generated three-dimensional cloud map and safety assessment report, thus ending the current multiphysics coupling analysis process that considers tidal hysteresis and deformation adaptive updates.

[0093] Application Examples: To aid in understanding the technical solution of this invention, an application example based on a deep foundation pit project of a river-crossing tunnel is provided below.

[0094] The excavation depth of the foundation pit is designed to be 25.4 meters, and the retaining structure adopts a 1.2-meter-thick underground continuous wall. The site strata, from top to bottom, are mainly silty clay, silty clay and slightly confined silt layer. The closest point of the outer side of the foundation pit to the river embankment is about 35.5 meters. This water area is significantly affected by tides and exhibits typical semi-diurnal tidal fluctuation characteristics.

[0095] The data acquisition module 10 installs a tide gauge on a fixed facility on the river surface and arranges multiple pore water pressure gauges along the depth direction in the soil layer within a range of 15 to 60 meters from the retaining structure outside the foundation pit. The burial depth is distributed between -12.0 meters and -32.0 meters. The system acquires a continuous 168-hour synchronous monitoring data sequence. The data processing module 20 performs Fourier transform and cross-correlation calculations on the time series data to extract the main frequency band phase difference of tidal fluctuations transmitted to different spatial measuring points. The calculation results show that the lag time is distributed between 1.8 hours and 4.7 hours. The data processing module 20 generates an initial three-dimensional infiltration lag time matrix containing the initial delay distribution of the loaded nodes in the entire domain based on ordinary kriging interpolation.

[0096] When the finite element simulation enters each excavation construction step, the topology evolution module 30 performs three-dimensional spatial Boolean difference set operation according to the excavation range, updates the node set of the excavation face, and uses the fast travel algorithm to solve the equation to obtain the new effective seepage path distance. The parameter update module 40 extracts the volumetric strain increment of the global mesh, calculates the local phase offset correction coefficient in combination with the seepage path distance, adaptively updates the lag time matrix and writes it into the memory interaction block. The boundary loading module 60 reads the memory data to complete the dynamic pore water pressure reconstruction. The solver engine 50 performs the Newton-Raphson incremental iterative calculation of the fluid-structure interaction equation, outputs pore pressure and strain distribution data, and determines the anti-seepage stability and safety state of each entity based on the hydraulic gradient and void ratio evolution parameters.

[0097] To verify the effectiveness of the method of the present invention, actual on-site monitoring data was extracted and compared with traditional numerical simulation schemes. Scheme 1 was set as the traditional finite element coupled calculation method. This scheme did not introduce tidal seepage lag correction, directly applied the river surface head elevation to the grid boundary synchronously, and did not update the feedback of soil deformation on seepage conduction delay in multi-condition analysis. Scheme 2 was set as the method provided by the present invention.

[0098] Error analysis was performed on the pore pressure data of three representative measuring points located on the outside of the foundation pit when the excavation reached the fourth excavation step (the excavation elevation reached -18.5 meters). These three measuring points were numbered P-1, P-2, and P-3, and were located at different depths of the slightly confined aquifer. The specific comparison is shown in the table below.

[0099] Table 1: Comparison of Absolute Errors in Predicting Peak Pore Water Pressure Extract the surface settlement monitoring data of the surrounding area during the phased excavation of the foundation pit, and compare the maximum settlement index under different working conditions. The specific comparison is shown in the table below.

[0100] Table 2: Comparison of Maximum Surface Settlement Around the Foundation Pit at Each Construction Stage Based on the data in Tables 1 and 2, and the appendix Figure 10 Appendix Figure 11 and attached Figure 12 It can be seen that the traditional finite element method, without considering the seepage conduction hysteresis and the dynamic evolution of porosity caused by soil deformation, seriously overestimates the peak pore water pressure, with the maximum calculation deviation reaching 28.57 kPa. The absolute error between the predicted value and the field measured value of the proposed method is stable within 3.5 kPa. The results show that the dynamic boundary reconstructed by the proposed method effectively filters out the false hydraulic gradient amplitude increment caused by synchronous loading and accurately restores the time lag and attenuation physical process of tidal load in complex porous media.

[0101] In the evolution of continuous unloading excavation of deep foundation pits, the surface settlement calculated by Scheme 1 is generally small. The main reason is that the static hydraulic boundary causes an abnormal rate of stress dissipation in the stratum, which fails to induce secondary consolidation deformation caused by the redistribution of seepage force.

[0102] The evolution trend of the maximum surface settlement calculated by the present invention is highly consistent with the measured results. The results show that the present method establishes a feedback closed loop between the stress field and the seepage delay time through matrix algebra operations, corrects the fluid-structure coupling stiffness around the foundation pit structure, and improves the accuracy of soil deformation prediction.

[0103] In summary, the above verification analysis shows that this scheme eliminates the systematic errors in the algorithm caused by the synchronous application of boundary head in conventional numerical simulations through the collaborative computation of multi-source temporal feature extraction, grid topology update and underlying memory interaction, providing high-precision data support for seepage stability assessment and disaster early warning of deep foundation pit projects near rivers.

[0104] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A smart construction simulation system for deep foundation pits in river-crossing tunnels with complex geological conditions, characterized in that, The system is deployed on a computer device, and the system includes: The data acquisition module is used to synchronously acquire physical environment time-series data, which includes a continuous time series of absolute elevation of the river tide level and a time series of initial pore water pressure on the outside of the foundation pit retaining structure. The data processing module is used to extract the phase difference from the physical environment time series data through cross-correlation operation, and to perform a spatial three-dimensional interpolation algorithm to map it to the loaded grid nodes in the three-dimensional initial finite element mesh model to generate the initial three-dimensional penetration hysteresis time matrix. The topology evolution module is used to identify the grid nodes of the excavation face based on the construction step sequence, calculate the shortest effective seepage path distance from the retained loaded grid nodes to the grid nodes of the excavation face along the grid skeleton, and generate a grid distance matrix. The parameter update module is used to extract the volumetric strain increment of the soil, calculate the local phase shift correction coefficient based on the grid distance matrix and the volumetric strain increment to generate a diagonal correction matrix, and calculate the local phase shift correction coefficient using the following formula: ； In the formula, For the calculation of the obtained loaded mesh nodes The local phase shift correction coefficient is set with a minimum cutoff threshold of 0.1 to prevent non-physical negative time delays caused by extreme soil expansion. The strain-time sensitivity coefficient characterizes the proportional relationship between the change in seepage lag time caused by the increase in strain per unit volume. It is calibrated based on the seepage consolidation test data during the engineering geological exploration stage, and its value ranges from 10.0 to 100.

0. For loaded mesh nodes The extrapolated volumetric strain increment is positive for tensile expansion and negative for compressive consolidation. For the natural constant Exponential function operators with base 0; Loaded grid nodes extracted from the grid distance matrix The shortest effective seepage path distance; The characteristic influence length characterizes the effective range of the impact of excavation disturbance on the seepage path. It is set according to the excavation depth of the foundation pit and ranges from 10.0 meters to 50.0 meters. The diagonal correction matrix is ​​constructed using all the calculated local phase offset correction coefficients as the main diagonal elements. The initial three-dimensional penetration hysteresis time matrix is ​​multiplied by the diagonal correction matrix to generate an updated hysteresis time matrix, and the updated hysteresis time matrix is ​​written into an independent memory block. The solver engine is used to construct the three-dimensional initial finite element mesh model containing the foundation pit structure and soil-water boundary, and output the volumetric strain increment in the iterative calculation. The boundary loading module is used to read the lag time from the memory interaction module within the time integration iteration step, and inject load data into the global equilibrium equation system of the solver engine to perform fluid-structure interaction calculations based on the calculated transient boundary conditions. The memory interaction module is allocated with the aforementioned independent memory blocks for boundary data caching interaction.

2. The intelligent construction simulation system for deep foundation pits of river-crossing tunnels in complex geological conditions as described in claim 1, characterized in that, The data acquisition module introduces a unified positioning system timestamp as a reference trigger signal to truncate and align the continuous time series of absolute elevation of the river surface tide with the initial pore water pressure time series on the time axis. The data acquisition module is equipped with a resampling unit, which performs cubic spline interpolation on the low-frequency data series to uniformly convert all input sequences into a set of equally spaced data points with the same discrete time step, generating a standardized digital matrix and outputting it to the data processing module.

3. The intelligent construction simulation system for deep foundation pits of river-crossing tunnels in complex geological conditions as described in claim 1, characterized in that, The data processing module performs a Fast Fourier Transform on the continuous time series of absolute elevation of the river tide and the initial pore water pressure time series, and filters out low-frequency trend terms and high-frequency environmental noise terms by setting a bandpass filter. In the time domain, it calculates the cross-correlation function of the filtered tide series and each pore water pressure series, and addresses the time offset corresponding to the peak value of the cross-correlation function as the phase difference of the main frequency band. The data processing module calls the three-dimensional ordinary kriging interpolation algorithm to calculate the interpolation weight coefficients of the loaded grid nodes for each discrete measuring point, and assembles the spatial interpolation results of the global nodes based on the interpolation weight coefficients to form the initial three-dimensional infiltration hysteresis time matrix.

4. The intelligent construction simulation system for deep foundation pits of river-crossing tunnels in complex geological conditions as described in claim 1, characterized in that, The topology evolution module extracts the set of geometric voxels of the soil mesh to be removed, performs a three-dimensional spatial Boolean difference operation in the finite element mesh coordinate system to extract the set of newly exposed excavation face mesh nodes; uses a fast traversal algorithm to solve the equations, extracts the permeability coefficient of each continuous solid element, takes the reciprocal to generate a discrete slow-degree field, traces the shortest arrival path radiating from the loaded mesh node to the set of excavation face mesh nodes along the finite element mesh skeleton, obtains and assembles the shortest effective seepage path distance of each node to generate the mesh distance matrix of the current construction step.

5. The intelligent construction simulation system for deep foundation pits of river-crossing tunnels in complex geological conditions as described in claim 1, characterized in that, The parameter update module extrapolates the volumetric strain increment at the integration point to each loaded grid node, and introduces the distance attenuation effect based on the effective seepage path for algebraic calculation.

6. The intelligent construction simulation system for deep foundation pits of river-crossing tunnels in complex geological conditions as described in claim 5, is characterized in that, The parameter update module constructs a preset analytical function for future tidal fluctuations based on historical tidal hydrological data, and uses the tidal harmonic analysis method to decompose tidal fluctuations into a superposition of multiple simple harmonic waves with fixed periods. The parameter update module converts the updated lag time matrix and the feature parameter set of the preset future tidal fluctuation analysis function into a continuous binary data stream, and directly writes the serialized binary data stream into the independent memory block in the physical running memory of the memory interaction module.

7. The intelligent construction simulation system for deep foundation pits of river-crossing tunnels in complex geological conditions as described in claim 6, characterized in that, At the time integration point, the boundary loading module reads the feature parameter set and the updated hysteresis matrix from the independent memory block through the memory interaction module; extracts the seepage conduction delay time data corresponding to the loaded grid nodes, subtracts the current absolute integration time from the seepage conduction delay time data to obtain the effective tidal fluctuation time, and substitutes it into the preset future tidal fluctuation analytical function to solve for the equivalent absolute head elevation at the location of each loaded grid node.

8. The intelligent construction simulation system for deep foundation pits of river-crossing tunnels in complex geological conditions as described in claim 7, is characterized in that, The boundary loading module introduces a truncation operator to limit the pore water pressure threshold. It calculates the difference between the absolute elevation of the equivalent head and the absolute vertical elevation of the loaded mesh node, compares this difference with the set minimum pore water pressure truncation threshold to select the maximum value, and then multiplies it by the volumetric unit weight of water to obtain the dynamic pore water pressure boundary value applied to the loaded mesh node. The dynamic pore water pressure boundary values ​​of all loaded mesh nodes are updated in the stiffness matrix and load vector of the finite element model to perform iterative solution.

9. The intelligent construction simulation system for deep foundation pits of river-crossing tunnels in complex geological conditions as described in claim 1, characterized in that, After completing the incremental iterative calculation of fluid-structure interaction, the solution engine performs a safety state determination, extracts pore water pressure data at each solid element to calculate the hydraulic gradient, extracts the volumetric strain data accumulated by the solid element in the current construction step, updates the current void ratio by combining it with the initial void ratio of the corresponding soil, and calculates the critical hydraulic gradient of the corresponding soil based on the relative density of soil particles and the current void ratio.

10. The intelligent construction simulation system for deep foundation pits of river-crossing tunnels in complex geological conditions as described in claim 9, characterized in that, The solver engine calculates the ratio of the critical hydraulic gradient to the hydraulic gradient as the anti-seepage stability safety factor, and introduces a small protection constant to prevent overflow when dividing by zero in the hydraulic gradient calculation term in the denominator; the anti-seepage stability safety factor is numerically compared with the preset engineering safety factor threshold. When the anti-seepage stability safety factor of a local area is less than or equal to the engineering safety factor threshold, the corresponding entity unit is determined to be in a seepage failure risk state, and the spatial coordinates and physical quantity distribution matrix of the risk unit are extracted to output the rendered safety state cloud map.

Images