Digital twinning system for whole process of pile foundation construction

Abstract

The invention relates to the technical field of digital twinning, in particular to a pile foundation construction whole process digital twinning system which comprises a twinning configuration module, a dynamic mapping module, a coupling analysis module and a twinning synchronization module. According to the method, a time axis driving model sequence is constructed based on space-time interpolation, burial depth coordinates, pouring time and segment numbers are dynamically bound, three-dimensional evolution mapping of a pile body model is achieved, axis coordinate deviation is combined with perpendicularity threshold extraction, virtual model coordinates are corrected in real time through point cloud registration, and a pump pressure value and settlement are fused to generate a pile sinking rate index. A bidirectional coupling relation between parameters and stratum response is established in combination with stratum counter-force and injection rate, the thickness of the model is adjusted according to a settlement rate difference value, the similarity of the injection rate and a counter-force curve is matched, a model parameter reverse updating mechanism is formed, and real-time intervention of construction abnormity, dynamic optimization of parameter coupling and model iteration correction are achieved. And the process controllability and the quality tracing precision are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of digital twin, and in particular to a digital twin system for the whole process of pile foundation construction. Background Art

[0002] The technical field of digital twin includes the construction of virtual mirrors for physical entities and their operation processes, real-time interaction, and state mapping. The core content of this technical field lies in using a cyber-physical fusion system to construct a virtual-real combined mapping relationship, relying on digital means such as the Internet of Things, big data, modeling and simulation, and edge computing to perform multi-dimensional and multi-level virtual reproduction of the real system. The overall framework of digital twin technology includes physical entity data acquisition, virtual model establishment, two-way interactive communication mechanism, and real-time mapping and predictive analysis functions driven by data, which are widely used in multiple industries such as industrial manufacturing, smart cities, and construction engineering to realize the modeling, tracking, evaluation, and optimization of physical processes.

[0003] Among them, the digital twin system for the whole process of pile foundation construction refers to a digital system applied to building foundation engineering, which combines multi-source sensing data acquisition technology, three-dimensional modeling methods, and construction process information analysis mechanisms to perform continuous modeling and data fusion on the pile foundation construction link. For the main construction stages such as pile hole formation, steel cage fabrication and hoisting, and concrete pouring, this system uses construction process video analysis, sensor data acquisition calibration, and three-dimensional space reconstruction methods to establish a time-sequential data model for the entire construction process; and based on this model, integrates the operation trajectories of construction equipment, records of engineering parameter changes, and environmental state acquisitions to construct a virtual construction mapping body reflecting the logical process of pile foundation construction, forming a model structure in which the whole process data corresponds dynamically to physical operations.

[0004] The existing technology relies on phased data acquisition and one-way model mapping, and the virtual model update lags behind the physical construction process. Multi-source sensor data is stored independently, and no time-sequential association framework for the steel cage segment number and pouring timestamp is established, resulting in insufficient spatial matching accuracy between the hoisting trajectory and the concrete pouring process. The verticality judgment uses a fixed threshold, lacking a coordinate correction mechanism based on real-time point cloud data, and it is easy to generate cumulative deviations when geological mutations occur during the pile hole formation process. The construction parameters and formation reaction force data do not form a dynamic weighted fusion relationship, and the pouring rate adjustment depends on manual experience, and there is an asynchronous delay between the model correction instruction and the physical operation. The existing virtual construction mapping body lacks a reverse feedback path, and the model thickness parameter cannot be dynamically iterated according to the settlement rate difference, and it is easy to cause axis deviation and structural stability risks under complex working conditions. Summary of the Invention

[0005] The purpose of the present invention is to solve the deficiencies existing in the prior art and propose a digital twin system for the whole process of pile foundation construction.

[0006] To achieve the above object, the present invention adopts the following technical solutions: A digital twin system for the whole process of pile foundation construction includes: A twin configuration module, which is used to generate a virtual pile body segmented model through the embedment depth coordinate sequence and the concrete pouring timestamp, call the spatio-temporal interpolation algorithm to construct a time-axis driven model sequence, bind the steel cage segment number and the timestamp, establish a dynamic mapping relationship, generate a virtual pile body segmented model and transfer it to the dynamic mapping module; A dynamic mapping module, which is used to extract the axis deviation value of the virtual pile body segmented model, calculate the Euclidean distance by unifying the coordinate units, call the verticality threshold to judge abnormal nodes, input the point cloud registration algorithm to correct the coordinates, generate a dynamic topological twin and transfer it to the coupling analysis module; A coupling analysis module, which is used to receive the dynamic topological twin, solve the theoretical value of the formation reaction force and calculate the pile sinking rate index, dynamically weighted fuse the injection rate and the measured value of the formation reaction force to generate perfusion coupling parameters, execute the dynamic weighted fusion algorithm to generate a construction feedback deviation index, trigger a twin correction instruction and transfer it to the twin synchronization module.

[0007] As a further solution of the present invention, the spatio-temporal interpolation algorithm synchronously updates the timestamp label of the model attribute node with the construction pouring progress by performing hash matching on the steel cage segment number and the timestamp; The verticality threshold is 5% of the pile diameter and is determined by verifying the on-site test pile data; The dynamic weighted fusion algorithm superimposes two types of parameters through normalized linear product, so that the error between the construction feedback deviation index and the actual deformation of the pile body is less than 3%.

[0008] As a further solution of the present invention, the twin configuration module includes: The coordinate time series processing sub-module obtains the embedment depth coordinate sequence and the concrete pouring timestamp, performs spatial hierarchical matching on the longitude, latitude, and elevation components of the coordinate points and the timestamp, calculates the difference between adjacent timestamp intervals based on the moving speed of the pouring machinery, and integrates the spatial nodes and the difference parameters to generate a pile body segmented coordinate set; The spatio-temporal interpolation modeling sub-module divides three-dimensional grid units based on the node distribution density of the pile body segmented coordinate set, uses the Lagrangian interpolation method to construct a time-axis increment sequence for the uncovered area, corrects the continuity of the grid vertex coordinates according to the timestamp label bound to the steel cage, and outputs a time-axis driven parameter group; The steel cage time binding sub-module extracts the steel cage segment number from the time-axis driven parameter group, maps the segment number to the segmented interval corresponding to the timestamp, and uses the hash matching algorithm to associate the segment number with the timestamp label of the model attribute node to generate a virtual pile body segmented model; The spatial hierarchical matching is realized by establishing a real-time association between the longitude and latitude coordinates and the GNSS positioning data of the construction machinery.

[0009] As a further solution of the present invention, the dynamic mapping module includes: The axis deviation extraction sub-module obtains the axis coordinate sequence of the virtual pile body segmentation model, decomposes the longitude, latitude, and elevation components of the model nodes, extracts the corresponding coordinate values of the theoretical axis, calculates the absolute value of the difference between the actual coordinates and the theoretical coordinates using the Euclidean distance formula, and generates the axis coordinate deviation value; The verticality verification sub-module, based on the axis coordinate deviation value, calls the axial verticality threshold, compares the deviation value item by item with the threshold in the order of nodes, marks the spatial index of the nodes whose deviation values exceed the axial verticality threshold, and generates an abnormal node identification set; The point cloud correction sub-module calls the point cloud registration algorithm to process the abnormal node identification set, extracts the coordinate offset of the abnormal nodes, matches the spatial distribution of the abnormal points and the theoretical point cloud through the iterative closest point algorithm, and introduces the pile body elastic deformation parameter to constrain the coordinate correction range, generating a dynamic topological twin body.

[0010] As a further solution of the present invention, the theoretical axis coordinate value is obtained by calculating the mean of the pile center coordinates on the design drawing and the construction layout point coordinates.

[0011] As a further solution of the present invention, the coupling analysis module includes: The formation reaction solution operator module receives the dynamic topological twin body, decomposes the pile diameter, pile length, and material elastic modulus in the twin body structure parameters, applies the Winkler foundation model to establish a linear equation between the formation reaction and the pile body displacement, solves the superposition theoretical value of the pile tip resistance and the side friction resistance, and generates the formation reaction theoretical value; The pile driving rate calculation sub-module calls the real-time pump pressure sensor data and the settlement amount measured by the laser rangefinder, converts the pump pressure value into the pile tip pressure component in kilonewtons and unifies it to the second-level time reference with the settlement amount, constructs an equilibrium equation between the pile driving resistance and the pump pressure based on the formation reaction theoretical value, and calculates the settlement amount increment per unit time using the forward difference method, and outputs the pile driving rate index; The dynamic fusion feedback sub-module, based on the pile driving rate index and the measured values of the pressure sensor, sets the injection rate weight coefficient as the reciprocal of the pump pressure change rate, and the formation reaction weight coefficient as the reciprocal of the measured value fluctuation variance, performs normalized linear product superposition on the two types of parameters through the dynamic weighted fusion algorithm, generates the construction feedback deviation index, and triggers the twin body correction instruction.

[0012] As a further solution of the present invention, the bed coefficient of the Winkler foundation model is determined by inverting the field static cone penetration test data.

[0013] As a further solution of the present invention, the system further includes: The twin synchronization module is used to adjust the thickness parameter of the virtual pile segment model based on the twin correction instruction, update the thickness according to the settlement rate difference, match the similarity of the injection rate and the curve shape of the reaction force measurement value, generate the updated twin parameters and synchronize them backward to the dynamic mapping module and the coupling analysis module.

[0014] As a further solution of the present invention, the updated twin parameters specifically refer to the adjusted pile segment thickness parameter and the material layer parameter; The shape similarity is calculated by calculating the sum of the squares of the slope differences of the two curves at adjacent time nodes, and screening the nodes with a difference exceeding 0.5 for weight ratio adjustment.

[0015] As a further solution of the present invention, the twin synchronization module includes: The thickness parameter adjustment sub-module calls the twin correction instruction, extracts the initial thickness parameter of the virtual pile segment model, and generates a thickness correction parameter according to the inverse proportional relationship between the thickness adjustment coefficient of the correction instruction and the pile driving rate index; The settlement rate update sub-module calculates the absolute value of the difference between the actual settlement rate and the theoretical rate based on the thickness correction parameter, and uses the cubic spline interpolation method to distribute the difference to the corresponding model nodes according to the time section length to generate a dynamic thickness update value; The curve shape matching sub-module obtains the time-domain sampling point sequences of the injection rate curve and the reaction force measurement value curve, calculates the sum of the squares of the slope differences of the two curves at multiple time nodes, screens the nodes exceeding the shape similarity threshold, and redistributes the weight ratio of the dynamic thickness update value to generate updated twin parameters.

[0016] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In the present invention, a time-axis driven model sequence is constructed based on the spatio-temporal interpolation algorithm, and the burial depth coordinates, perfusion timestamps and steel cage segment numbers are dynamically bound to realize the continuous evolution mapping of the pile segment model in three-dimensional space. The extraction of the axis coordinate deviation is combined with the verticality threshold judgment mechanism, and the real-time coordinate correction of the virtual model is carried out through the point cloud registration algorithm to meet the immediate calibration requirements for sudden offsets during construction. The real-time pump pressure value and settlement data are fused to generate the pile driving rate index, and the dynamic weighted parameters of the formation reaction force measurement value and the injection rate are synchronously fused to establish a two-way coupling relationship between the construction parameters and the formation response. The model thickness parameter is adjusted based on the settlement rate difference, and the curve shape similarity of the injection rate and the reaction force measurement value is inversely matched to form a reverse synchronous update mechanism for the virtual model parameters, ensuring a dynamic closed-loop of physical operations and data feedback. This processing logic breaks through the traditional single-stage static modeling limitation, transforms the discrete construction links into continuous dynamic mapping, realizes real-time intervention in construction anomalies, dynamic optimization of parameter coupling and reverse iterative correction of the model, and significantly improves the controllability of the construction process and the accuracy of quality traceability. Description of the Drawings

[0017] Figure 1 This is the system flow chart of the present invention; Figure 2 This is the flow chart of the twin configuration module of the present invention; Figure 3 This is the flow chart of the dynamic mapping module of the present invention; Figure 4 This is the flow chart of the coupling analysis module of the present invention; Figure 5 This is the flow chart of the twin synchronization module of the present invention. Detailed Embodiments

[0018] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0019] In order to make the purpose, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention, rather than to limit the present invention.

[0020] In the description of the present invention, it should be understood that the orientation or positional relationship indicated by the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. is based on the orientation or positional relationship shown in the accompanying drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation of the present invention. In addition, in the description of the present invention, the meaning of "a plurality" is two or more, unless otherwise specifically defined.

[0021] Embodiment 1

[0022] Please refer to Figure 1 , a digital twin system for the whole process of pile foundation construction includes: A twin configuration module, which is used to generate a virtual pile body segmented model through the burial depth coordinate sequence and the concrete pouring timestamp, call the spatio-temporal interpolation algorithm to construct a time-axis driven model sequence, bind the steel cage segment number and timestamp to the model attributes, establish a dynamic mapping relationship between the segmented model and the construction progress, generate a virtual pile body segmented model, and transfer it to the dynamic mapping module; The dynamic mapping module is used to extract the axis coordinate deviation value based on the virtual pile segment model, unify the elevation component and the plane coordinates into meters, calculate the Euclidean distance, call the verticality threshold to judge abnormal nodes, input the point cloud registration algorithm to correct the coordinates, generate a dynamic topological twin and transmit it to the coupling analysis module; The coupling analysis module is used to perform coupling analysis after receiving the dynamic topological twin, calculate the theoretical value of the formation reaction force based on the twin structure parameters, calculate the pile driving rate index by combining the real-time pump pressure value and the settlement amount per unit time, and allocate the injection rate weight coefficient and the formation reaction force weight coefficient according to the reciprocal of the pump pressure change rate and the reciprocal of the measured value fluctuation variance. Perform dynamic weighted fusion on the injection rate and the formation reaction force measured value to generate perfusion coupling parameters, input the dynamic weighted fusion algorithm to generate a construction feedback deviation index, trigger a twin correction instruction and transmit it to the twin synchronization module; The twin synchronization module is used to adjust the thickness parameter of the virtual pile segment model based on the twin correction instruction, update the thickness according to the settlement rate difference, match the similarity of the curve shapes of the injection rate and the reaction force measured value, generate the updated twin parameters and synchronize them back to the dynamic mapping module and the coupling analysis module.

[0023] The virtual pile segment model includes the steel cage segment number and the concrete perfusion timestamp. The dynamic topological twin is specifically the corrected axis coordinates and the continuous pile segment topological relationship. The construction feedback deviation index includes the pile driving rate index and the perfusion coupling parameters. The updated twin parameters specifically refer to the adjusted pile segment thickness parameter and the material layer parameter.

[0024] The spatio-temporal interpolation algorithm synchronizes the timestamp label of the model attribute node with the construction perfusion progress by hashing and matching the steel cage segment number and the timestamp; The verticality threshold is 5% of the pile diameter and is determined by verifying the on-site test pile data; The dynamic weighted fusion algorithm superimposes two types of parameters through normalized linear multiplication, making the error between the construction feedback deviation index and the actual deformation of the pile body less than 3%; The shape similarity is calculated by calculating the sum of the squares of the slope differences of two curves at adjacent time nodes, and screening the nodes with differences exceeding 0.5 for weight ratio adjustment.

[0025] Please refer to Figure 2 , the twin configuration module includes: The coordinate time series processing sub-module obtains the buried depth coordinate sequence and the concrete perfusion timestamp, performs spatial hierarchical matching on the longitude, latitude, and elevation components of the coordinate points and the timestamp, calculates the difference between adjacent timestamp intervals based on the moving speed of the perfusion machinery, and integrates the spatial nodes and the difference parameters to generate the pile segment coordinate set; The coordinate time series processing sub-module obtains the buried depth coordinate sequence and the concrete pouring timestamp. The specific execution process is as follows: First, the system collects the original monitoring data from the sensor network installed at the construction site (such as the GNSS high-precision positioning module and depth encoder integrated on the drilling equipment, or the multi-point displacement and pressure sensors pre-fixed on the steel reinforcement cage). Taking a specific engineering pile P001 as an example, the designed pile length of this pile is 30 meters and the designed diameter is 0.8 meters. During the drilling and steel reinforcement cage lowering process, a series of three-dimensional coordinate points representing the pile hole or the steel reinforcement cage at different depths will be recorded. , where is the longitude, is the latitude, is the elevation (or the depth relative to the unified reference plane). These coordinate point data, together with the corresponding acquisition time, are accurately recorded.

[0026] During the concrete pouring stage, the concrete pumping control system will record the key pouring operation timestamps in real time, and these timestamps are accurate to the second. When the concrete liquid level rises to a predetermined or monitored specific depth interval (perceived by the liquid level sensor installed at the end of the conduit or on the steel reinforcement cage), the system matches the spatial position of this depth interval with the corresponding pouring completion timestamp. Specifically, the GNSS positioning system of the construction machinery (such as the concrete pump truck or drill rig) provides its planar position information (such as the coordinates of the pump truck near pile P001 during operation). ), and this information is used to confirm that the current pouring operation is for pile P001. The system will assign specific pouring timestamps to the corresponding vertical segments of the pile according to the theoretical pumping volume of the concrete and the actually monitored rising speed of the concrete liquid level in the pile hole.

[0027] Subsequently, based on the controlled lifting speed of the pouring machinery (such as the concrete conduit) in the pile hole (this speed is set according to the requirements of the concrete pouring process, for example, controlled between 0.15 meters and 0.3 meters per minute, and here it is set as the average vertical lifting speed). ), and the actually monitored pouring completion time of each segment, calculate the interval difference and between adjacent effective pouring timestamps (corresponding to the moments when the pouring of the pile segments from depth to is completed). Let (corresponding to the completion of pouring in the range of depth -0.5m to -1.5m), (corresponding to the completion of pouring in the range of depth -1.5m to -2.5m), then the time interval difference .

[0028] This difference parameter and the corresponding spatial node set (by depth and and the pile segments defined by the planar coordinates of this depth segment). Finally, the system integrates this spatial node information with the time difference parameters to generate a structured set of pile segment coordinates. Each element in this set contains the starting and ending three-dimensional coordinates of a pile segment (e.g., Segment 1: starting from , ending at ), as well as the perfusion completion timestamp precisely associated with it and the perfusion time consumption of this segment . For pile P001, it will be carefully divided into multiple such segments, each with its unique spatio-temporal attributes.

[0029] Table 1: Monitoring Data of Pile Embedded Depth Coordinates and Perfusion Time of Pile Foundation P001 (Partial)

[0030]

[0031] As shown in Table 1, this table details the geographical coordinate range, elevation range, concrete perfusion completion timestamp of each segment, and the perfusion time consumption of each segment for the first five segments of pile foundation P001. These precisely recorded data are the basis for subsequent construction of a high-precision digital twin model.

[0032] The spatio-temporal interpolation modeling sub-module divides three-dimensional grid cells based on the node distribution density of the set of pile segment coordinates, constructs an incremental sequence of the time axis using Lagrange interpolation for the uncovered areas, corrects the continuity of the grid vertex coordinates according to the timestamp tags bound to the steel reinforcement cage, and outputs a set of time axis driving parameters; The spatio-temporal interpolation modeling sub-module divides three-dimensional grid cells based on the node distribution density of the set of pile segment coordinates generated in the previous step. The system analyzes in detail the set of pile segment coordinates of pile P001, which contains the precise starting and ending depths of each segment (continuing the data in Table 1, e.g., for segment P001-S1, the depth ranges from -0.5m to -1.5m) and the corresponding perfusion completion timestamps. Based on the geometric dimensions of these segments (pile diameter , segment length usually 1m or less) and the data point density, the basic size of the three-dimensional grid cells is determined. To ensure a full representation of the pile shape, the grid cell size is set to .

[0033] In actual construction monitoring, if due to temporary failure of sensors or interruption of data transmission, some pile segments (such as between depths -5.5m and -6.5m) lack dense sensor data, and there are only the perfusion timestamps at both ends of this segment (i.e., at -5.5m and -6.5m) (corresponding to the end of segment P001-S5) and If it is the next valid recording point, the system uses Lagrange interpolation to construct the incremental sequence of these uncovered regions on the time axis. During specific execution, based on the known two end timestamps , and the length of this section , and referring to the average perfusion rate of the adjacent covered region (such as the P001-S5 section) (calculated to be approximately 0.00328 m / s from its length of 1.0 m and time consumption of 305 seconds), to estimate any virtual node inside this 1-meter uncovered region (located between -5.5 m and -6.5 m, and the depth from -5.5 m is )'s approximate perfusion completion timestamp . If a timestamp of a virtual node is to be inserted at -6.0 m, this point is located in the middle of this 1-meter section ( ), and its timestamp is estimated to be . Through this method, corresponding time series data is generated for grid cells with sparse or missing data.

[0034] Subsequently, the system retrieves the pre-recorded and stored time-stamp tags of the steel reinforcement cage. These tags are accurately bound to specific nodes of the steel reinforcement cage during the segmented fabrication and hoisting of the steel reinforcement cage (such as the top, middle reinforcement ring, and bottom of each section of the steel reinforcement cage). The steel reinforcement cage of P001 pile has three sections, each 10 meters long. The hoisting-in-place timestamp of the top of the first section of the steel reinforcement cage (covering 0 m to -10 m) is . The system uses these actual time-stamp tags of the steel reinforcement cage to correct the continuity deviation of the grid vertex coordinates in the time series that may be introduced due to interpolation or other modeling processes. If the perfusion timestamp of a grid vertex obtained by interpolation has a difference outside a set allowable deviation from the actual timestamp when the steel reinforcement cage is lowered or covered by concrete at this position, then it is corrected based on the actual time record of the steel reinforcement cage. The setting of

[0035] After the above processing, the system finally outputs a set of time-axis-driven parameter groups. This parameter group contains in detail the timestamp information of each vertex of the three-dimensional grid cell, as well as its corresponding precise spatial coordinates, laying a solid foundation for subsequent dynamic analysis.

[0036] The steel cage time binding sub-module extracts the steel cage segment numbers from the time-axis-driven parameter group, maps the segment numbers to the segmented intervals corresponding to the timestamps, and uses the hash matching algorithm to associate the segment numbers with the timestamp tags of the model attribute nodes to generate a virtual pile body segmented model; The steel cage time binding sub-module extracts the segment numbers of the steel cage from the time-axis-driven parameter group generated by the aforementioned "space-time interpolation modeling sub-module". This time-axis-driven parameter group contains the precise spatial coordinates of each vertex of the pile body three-dimensional grid and the corresponding timestamps . The detailed segmentation information of the steel cage of pile No. P001 is pre-stored in the system: the total length is 30 meters, divided into (0m to -10m), (-10m to -20m), (-20m to -30m) for three sections. When each section of the steel cage is hoisted into the hole, the accurate hoisting completion time of its key positions (such as the top and bottom) is recorded and associated with the segment number. For example, the hoisting completion time of segment (i.e., the time when its bottom reaches a depth of -10m) is , and the hoisting completion time of segment (bottom reaches -20m) is .

[0037] The system filters out the grid vertices in the time-axis-driven parameter group that spatially coincide with or are adjacent to the physical positions of each section of the steel cage, and extracts their corresponding timestamps. These timestamps are then matched with the pre-stored construction time windows of each section of the steel cage (the time period from the hoisting completion of this section to the time when it is completely covered by concrete). For example, the grid cells in the virtual pile body model with a depth range of -0.5m to -9.5m, the associated perfusion timestamp sequence (starting from 10:02:30 as shown in Table 1) are all later than the hoisting completion time of segment (09:15:00), and earlier than the time when segment starts to have a significant impact (i.e., the hoisting of and its subsequent perfusion). Therefore, these grid cells located in the depth range of -0.5m to -9.5m are assigned the steel cage segment number

[0038] Next, the system uses the hash matching algorithm to efficiently associate the extracted steel cage segment numbers (such as Precisely associate with the timestamp tags of the corresponding attribute nodes in the virtual pile body model. The specific execution process of hash matching is as follows: Generate a unique hash value (e.g., hash("RC_1") gets 0x5D8F1A) for each steel cage segment number (such as "RC_1"). For each model attribute node in the time-axis-driven parameter group (carrying spatial coordinates and perfusion timestamps ), the system determines which steel cage segment's actual influence time interval and spatial range this node belongs to based on its and its spatial depth. For example, for model node N789, whose spatial position is at -15.5m and whose perfusion timestamp is '2025-05-12 10:55:00', this time point and spatial position clearly correspond to the stage where the steel cage segment (-10m to -20m, with the hoisting completion time of 09:50:00) is being concreted, then the identifier of node N789 is associated with and stored with the hash value hash("RC_2") (such as 0x9B3C7E) of . This fast lookup and association mechanism based on the hash table enables the system to quickly and accurately assign the correct steel cage segment timestamp tags to a large number of attribute nodes that make up the virtual pile body model.

[0039] Through the above refined processing, a virtual pile body segmented model with clear segmentation information and accurate time attributes is finally generated. This model not only accurately reproduces the three-dimensional geometric shape of the pile body, but more importantly, it embeds the chronological order of the formation of each component of the pile body, as well as the firm corresponding relationship between these parts and the corresponding steel cage segments. The spatial hierarchy matching is achieved by establishing a real-time association between the longitude and latitude coordinates and the construction machinery GNSS positioning data.

[0040] Please refer to Figure 3 , the dynamic mapping module includes: The axis deviation extraction sub-module obtains the axis coordinate sequence of the virtual pile body segmented model, decomposes the longitude, latitude, and elevation components of the model nodes, extracts the corresponding coordinate values of the theoretical axis, calculates the absolute value of the difference between the actual coordinates and the theoretical coordinates using the Euclidean distance formula, and generates the axis coordinate deviation value; The axis deviation extraction sub-module is responsible for precisely obtaining the coordinate point sequence of the actual central axis of the pile body from the virtual pile body segmented model generated by the "steel cage time binding sub-module". This sequence is composed of the geometric center points of the pile body model at different monitoring depths (elevations . For pile P001, at each horizontal cross-section at depth (taking a cross-section every 1 meter), the system calculates the precise coordinates of the centroid of its cross-section , thus obtaining a series of discrete points on the actual axis: .

[0041] Subsequently, the system decomposes the coordinate components of these actual axis nodes. If the original coordinates are geographical coordinates (latitude and longitude), they are first uniformly converted to the values in the project construction coordinate system. Here, the actual axis coordinates of point are represented as . Then, the coordinate values of the theoretical axis at the corresponding depth are extracted. The theoretical axis represents the centerline position of the pile under ideal design conditions, and its coordinates are clearly given in the design drawings, that is, the designed center coordinates of the pile top and the designed center coordinates of the pile bottom . For vertical piles, the plane coordinates of the theoretical axis at any depth are constant. This theoretical plane coordinate is calculated by taking the average of the pile center coordinates in the design drawings (the designed pile center coordinates of pile P001 are ) and the pile position center coordinates after on-site precise construction layout review (the measured layout pile center coordinates are ) to reduce the error from a single source: ; Therefore, the theoretical axis plane coordinates of pile P001 are .

[0042] Then, the Euclidean distance formula is used to calculate the offset distance on the horizontal plane between the actual axis point and the theoretical axis point at the same depth , that is, the axis coordinate deviation value at this depth. The calculation formula is: ; at the depth , the coordinates of the actual axis point of pile P001 obtained through model calculation are . Then, the axis coordinate deviation value at this depth is calculated as follows: the difference ; the difference ; the sum of squares ; that is, 67.082 mm.

[0043] The system performs this deviation calculation for each actual axis point of pile P001 along its entire length (from -0.5 m to -29.5 m, a total of 30 calculation sections), and finally generates a sequence The advantage of this formula is that by taking the square root of the sum of the squares of the differences between the actual coordinate components and the theoretical coordinate components, it can objectively and directly quantify the degree to which the axis of the pile deviates from its ideal design position at each depth, providing a key quantitative basis for subsequent verticality verification and point cloud correction. The calculation result (i.e., 67.082 mm) indicates that at the depth , the horizontal offset of the actual axis of Pile P001 compared to its theoretical axis is 67.082 mm. This value will serve as a key input for the subsequent verticality verification sub-module.

[0044] Based on the axis coordinate deviation value, the verticality verification sub-module calls the axial verticality threshold, compares the deviation value item by item with the threshold in the order of nodes, marks the spatial index of the nodes where the deviation value exceeds the axial verticality threshold, and generates a set of abnormal node identifiers; Based on the sequence of axis coordinate deviation values generated by the previous "axis deviation extraction sub-module", the verticality of the pile is verified for compliance. Each element in this sequence accurately represents the horizontal offset between the actual axis and the theoretical axis of the pile at a specific depth . Taking the deviation value calculated for Pile P001 at a depth of -10.5 m as an example.

[0045] The system calls the preset allowable deviation threshold for axial verticality . The setting of this threshold strictly follows the current national "Technical Code for Building Pile Foundations" (such as JGJ94) and the specific requirements in the project design documents. For Pile P001 (with a diameter , and a pile length ), the code usually stipulates that the allowable value for the deviation of the pile top of a vertical pile is a certain multiple of the pile diameter or an absolute value, and the allowable value for the deviation at any cross-section of the pile body is a specific proportion of the pile length. Here, for the verticality of any cross-section of the pile body, the allowable deviation threshold is set to the smaller value between 1 / 350 of the pile length and 100 mm. Calculation process: The allowable deviation ratio of the pile length . The upper limit of the absolute allowable deviation . Then . This threshold The setting basis is as follows: The pile length ratio limit of 1 / 350 is obtained based on long-term engineering practice and theoretical analysis of pile foundation bearing capacity and structural interaction, and is used to control the additional stress caused by excessive bending of the pile shaft; while the absolute upper limit of 100 mm takes into account the general level of construction technology and the tolerance of the superstructure to pile position deviation. To ensure the rationality of this threshold, the following verification was carried out: 15 full-scale model piles under similar conditions to P001 pile were selected for static load tests. The maximum deflection of the pile shaft of these model piles ranged from 50 mm to 150 mm (manufactured at intervals of 10 mm for each group). The test results (partial data as shown in Table 2) show that when the maximum deflection of the pile shaft exceeds 1 / 350 of the pile length (about 85.7 mm), the horizontal bearing capacity of the pile begins to show non-linear decline, and the critical point where the decline amplitude exceeds 10% is close to this value; when the deflection approaches 100 mm, the bearing capacity decline amplitude of some piles exceeds 18%, and the stress concentration phenomenon of the pile shaft intensifies. Considering safety and economy comprehensively, is selected as the control standard.

[0046] Table 2: Test data of model pile bearing capacity under different pile shaft deflections (partial)

[0047]

[0048] As shown in Table 2, this table lists the deflection conditions of some model piles and their corresponding horizontal ultimate bearing capacities. The data shows that when the deflection exceeds 85.7 mm (such as pile M-6), the bearing capacity decline rate begins to increase significantly, verifying the rationality of the threshold selection.

[0049] The system will use the previously generated 30 axial coordinate deviation values in depth order and compare them item by item with this threshold The comparison judgment condition is If is greater than , it is determined that the verticality of the pile at this node does not meet the specification requirements. For the deviation value at a depth of -10.5 m, since , the verticality of this node is qualified. If the calculated deviation value at a depth of -18.5 m, then because , the verticality of this node is unqualified. The system will mark the node (i.e., the node at a depth of -18.5 m) where this deviation value exceeds the axial verticality threshold and record its spatial index (for example, the serial number of this node in the axial coordinate sequence is 19, and its three-dimensional spatial coordinates are ). The system collects the spatial indices of all nodes determined to be unqualified in perpendicularity, and finally generates a set of abnormal node identifiers. If the test results of P001 pile show that the axis deviations at depths of -15.5m, -18.5m, and -22.5m all exceed 85.71mm, the set of abnormal node identifiers will contain the detailed information of these three depth nodes.

[0050] The point cloud correction sub-module calls the point cloud registration algorithm to process the set of abnormal node identifiers, extracts the coordinate offsets of abnormal nodes, matches the spatial distribution of abnormal points with the theoretical point cloud through the iterative closest point algorithm, and introduces the pile body elastic deformation parameter to constrain the coordinate correction range to generate a dynamic topological twin. The point cloud correction sub-module calls the point cloud registration algorithm to process the set of abnormal node identifiers generated by the "perpendicularity verification sub-module" in the previous step. This set contains the spatial indices or precise coordinates of the pile body nodes in P001 pile whose actual axis deviations exceed the preset perpendicularity threshold . Suppose the set of abnormal nodes contains node J (located at a depth of -18.5m, with actual coordinates , and its axis deviation ) and node K (located at a depth of -22.5m, with actual coordinates , and its axis deviation ).

[0051] The system first accurately extracts the point cloud data of these abnormal nodes (J, K, etc.) and a certain neighborhood range around them from the initially constructed three-dimensional point cloud model of the pile body (this model is generated by processing multi-source sensor data during construction). At the same time, the system obtains the points corresponding to the theoretical axis positions of these abnormal nodes (the corresponding point of node J on the theoretical axis is , and its coordinates are ). Based on this, the current coordinate offset vector of each abnormal node is calculated, mainly focusing on its horizontal offset, such as the horizontal offset vector of node J.

[0052] Next, the system uses the iterative closest point (ICP) algorithm to accurately match the spatial distribution of the actual point cloud segment containing these abnormal regions with the pile body design model (or its theoretical point cloud representation). The specific implementation steps of the ICP algorithm are as follows: 1. For each point in the abnormal point cloud segment, search for and determine its unique nearest neighbor point in the point cloud of the theoretical design model. 2. Based on the set of established corresponding point pairs, calculate an optimal rigid transformation parameter (including a rotation matrix and a translation vector). This transformation aims to minimize the root mean square error (RMSE) between the abnormal point cloud after transformation and its corresponding points in the theoretical point cloud, so that the two sets of point clouds coincide to the greatest extent. 3. Apply the calculated transformation Applied to the entire abnormal point cloud segment. 4. Repeat steps 1 to 3 until the change in transformation parameters between two iterations is less than the preset convergence threshold (the convergence threshold is set to 0.1 mm, indicating that the average point position correction accuracy reaches 0.1 mm), or the number of iterations reaches the preset maximum number of iterations (the maximum number of iterations is set to 100 times). The setting of this convergence threshold and the maximum number of iterations is based on a trade-off between algorithm efficiency and final registration accuracy. Through multiple registration tests on standard geometric bodies and observing the RMSE decline curve, parameters at the balance point of computational cost and accuracy gain are selected.

[0053] During the process of the ICP algorithm performing matching and correction, the system introduces the elastic deformation parameters of the pile body material as key constraints. These parameters include the elastic modulus of C35 concrete used for P001 pile and Poisson's ratio . The purpose of introducing these parameters is to ensure that the scope and method of coordinate correction conform to the mechanical properties of the material, and to avoid generating unrealistic correction results that would cause excessive deformation of the pile body model or abnormal local stress concentration. This means that the transformation calculated by the ICP algorithm cannot cause the pile body to deform beyond its elastic limit. If directly applying the calculated transformation by ICP would result in the local curvature radius of the pile body being less than the minimum allowable curvature radius determined by the flexural performance of the material (calculated based on the pile diameter and reinforcement, set to 250 times the pile diameter, that is ), or the strain generated by the relative displacement between any two points exceeds the allowable strain value of the concrete (such as 0.002, this value is determined according to the concrete compressive strength grade and safety factor), then the system will adjust the correction amount of the ICP. The adjustment method is: on the basis of maintaining the offset correction direction, recalculate the allowable correction amplitude according to the elastic constraint conditions, or proportionally reduce the original displacement correction vector to ensure that the deformation of the corrected pile body is still within the elastic range.

[0054] By performing such registration and correction operations strictly constrained by material elastic parameters on the point cloud data of all identified abnormal areas, the system finally generates a dynamic topological twin. This twin is closer to the design intention of the pile body in terms of macroscopic geometric shape (i.e., reducing the axis deviation), and at the same time scientifically retains the necessary local actual deviation information that conforms to physical laws and material characteristics, enabling it to more realistically reflect the actual occurrence state of the pile body.

[0055] The theoretical axis coordinate value is obtained by calculating the average of the pile center coordinate on the design drawing and the construction lofting point coordinate.

[0056] Please refer to Figure 4 , the coupling analysis module includes: The formation reaction force solution operator module receives the dynamic topological twin, decomposes the pile diameter, pile length, and elastic modulus of the material in the twin structure parameters, applies the Winkler foundation model to establish a linear equation between the formation reaction force and the pile body displacement, calculates the superposition theoretical value of the pile tip resistance and the side friction resistance, and generates the theoretical value of the formation reaction force. The formation reaction force solution operator module receives the dynamically topological twin model that has been finely corrected and optimized and output by the "point cloud correction sub-module". This model accurately represents the actual three-dimensional geometric shape of P001 pile and its spatial position in the soil mass in the form of a three-dimensional grid or dense point cloud.

[0057] The system first decomposes and extracts the key parameters necessary for accurate calculation of the formation reaction force from the structure parameter database of this twin model. These parameters mainly include: pile diameter , which is the equivalent average diameter calculated by back-calculating according to the corrected model at different depths. For P001 pile, its overall average corrected pile diameter is taken as ; pile length , which is the accurate length from the designed elevation of the pile top to the actual depth of the pile bottom, taken as ; and the elastic modulus of the pile material (C35 concrete) .

[0058] Next, the system applies the basic theory of the Winkler foundation beam model to establish the mathematical relationship between the formation reaction force and the pile body displacement. The Winkler model simplifies the foundation around the pile into a series of discrete elastic springs distributed along the pile shaft, each of which generates a reaction force proportional to its own compression displacement (i.e., the displacement of the pile body at this point). The relationship is , where is the formation reaction force per unit area at depth , is the lateral or vertical displacement of the pile body at this depth, and is the foundation reaction coefficient (also known as the bedding coefficient) that varies with depth. The accurate determination of the bedding coefficient is crucial for the calculation results. It is estimated through analyzing the detailed geotechnical investigation report of the construction site of P001 pile, especially the in-situ test data (such as the number of blows in the standard penetration test SPT, the tip resistance and the sidewall friction resistance of the static cone penetration test CPT, and the results of the pressuremeter test PMT) and combining the physical and mechanical property indexes of each soil layer, through empirical formulas or numerical inversion methods. The main soil layers penetrated by P001 pile and their corresponding bedding coefficient values are as follows: -0m to -8m: silty clay, average silty clay with silt, average fine sand, average These The value setting is based on the analysis of the CPT data of the corresponding soil layers in the exploration report. For example, for the fine sand layer, its average cone tip resistance , according to the empirical relationship (where is the empirical coefficient, and the value range for sandy soil is 0.8 - 1.5, and here it is taken as 1.0), estimate . In actual applications, more refined stratification and more complex inversion models will be adopted. Here, for simplicity of illustration, the stratified average value is used.

[0059] The system uses these parameters to solve for the pile shaft displacement and internal force through the finite element method or the finite difference method, and then calculates the ultimate end resistance and the total ultimate skin friction resistance of the pile shaft. For the end resistance of the pile, its theoretical value is calculated based on the bearing capacity characteristics of the end bearing layer of the pile and the end area of the pile. For the total skin friction resistance of the pile shaft, it is the integral (or piecewise accumulation) of the unit side friction resistance provided by each soil layer along the pile shaft length. The unit side friction resistance is closely related to the properties of the soil at that depth, the effective stress between the pile and the soil, the relative displacement, and the bedding coefficient . After calculation, the theoretical ultimate end resistance of pile P001 and the total theoretical ultimate skin friction resistance of the pile shaft are obtained. Therefore, the theoretical value of the superimposed total formation ultimate reaction force (i.e., the vertical ultimate bearing capacity of the pile) is . This 7150 kN is the theoretical ultimate value of the formation reaction force generated under the current conditions.

[0060] The pile driving rate calculation sub-module calls the real-time pump pressure sensor data and the settlement measured by the laser rangefinder, converts the pump pressure value (MPa) into the pile end pressure component in kN units and unifies the settlement (mm) to the second-level time reference, constructs the balance equation of the pile driving resistance and the pump pressure based on the theoretical value of the formation reaction force, and uses the forward difference method to calculate the settlement increment per unit time, and outputs the pile driving rate index; The pile driving rate calculation sub-module calls the pump pressure sensor data monitored in real time during the construction process and the pile settlement measured by the laser rangefinder (or high-precision static level). The system real-time collects two types of key data streams from the database of the construction monitoring platform: The first is the pressure sensor data from the concrete pumping system (or in the case of static pile driving construction, from the hydraulic system), and this data is recorded in megapascals (MPa). Suppose at a specific moment, the pump system cylinder pressure reading is . The second is the real-time pile settlement measured by the laser rangefinder installed on the pile top reference point or the fixed reference object, and this data is in millimeters (mm). At the same moment, the cumulative settlement of the pile top of pile P001 is .

[0061] The system first performs unit conversion and strict unification of the time reference on the collected raw data. Convert the pump pressure value (MPa) to the equivalent pile top driving force component in kilonewtons (kN) . This conversion needs to consider the effective area of the hydraulic cylinder and the mechanical and pressure transmission efficiency coefficient . is a known parameter determined according to the model of the static pile driver used, which is here is a comprehensive efficiency coefficient, reflecting the loss from the cylinder pressure to the effective pressure actually applied to the pile top. Its value is determined through equipment calibration experiments. The calibration experiment process is as follows: Before pile foundation construction, use a standard force sensor to directly measure the actual pile driving force output by the pile driver under different cylinder pressures. At least 5 groups of tests with different pressure levels are carried out, and each group is repeated 3 times, and the average value is calculated . Through this calibration is determined to be 0.88. Unit conversion rule: 1MPa = 1000kN / m². Therefore, the equivalent pile top driving force . Calculation: .

[0062] (Note: The 132.0kN here is the total driving force applied. The effective force actually used to overcome the formation resistance and cause settlement also needs to consider factors such as the self-weight of the pile body. Here, it is simplified that the equivalent driving force directly balances with the formation reaction force or causes movement). At the same time, convert the settlement (mm) to meters (m) uniformly, that is . All data points are accurately aligned to the unified second-level time reference .

[0063] Subsequently, based on the theoretical limit value of the formation reaction force output by the "formation reaction force solution operator module" (this value represents the total ultimate resistance that the soil can provide at the current total depth of the pile. At each moment during actual pile driving, the actual resistance provided by the formation is less than or equal to this limit value and changes with the development of settlement), and the self-weight of the pile body calculated in real time (increasing with the increase of pile length), construct the force balance equation or motion equation during pile driving. In uniform pile driving or quasi-static analysis, the driving force is approximately equal to the total resistance. In the process of dynamic pile driving, the instantaneous resultant force (if the self-weight of the pile is considered as a favorable factor for pile driving, it is ) drives the pile body to move. The system uses the forward difference method to calculate the settlement increment per unit time, that is, the pile driving rate. The specific operation is as follows: At two adjacent sampling time points and (time interval , the sampling interval of this system is fixed at ), the cumulative settlement amounts of the pile top are measured to be and . Then the average pile driving rate within this time period is calculated as . If at o'clock , at the subsequent o'clock, is measured. Then the average pile driving rate within these 2 seconds. The calculated value is the current pile driving rate index of the output.

[0064] The dynamic fusion feedback sub-module, based on the pile driving rate index and the measured values of the pressure sensor, sets the injection rate weight coefficient as the reciprocal of the pump pressure change rate, and the formation reaction force weight coefficient as the reciprocal of the measured value fluctuation variance. Through the dynamic weighted fusion algorithm, it performs normalized linear product superposition on the two types of parameters to generate a construction feedback deviation index and trigger the twin body correction instruction; The dynamic fusion feedback sub-module is based on the real-time pile driving rate index (the current value is ) output by the previous "pile driving rate calculation sub-module" and the real-time equivalent pile top driving pressure directly measured by the pressure sensor (represented by the pump pressure here, the current value is , and its change trend indirectly reflects the change characteristics of the formation reaction force or the stability of construction control) for comprehensive analysis.

[0065] The system sets two types of key weight coefficients: the weight coefficient of the injection rate (or construction efficiency, represented by the pile driving rate here), and the weight coefficient of the formation reaction force response (represented by the stability of the pump pressure ). The injection rate weight coefficient is set as the normalized form of the reciprocal of the change rate of the pile driving rate within a past period of time (such as 60 seconds). Specific calculation: First, calculate the pile driving rate every 2 seconds within the past 60 seconds to obtain 30 rate values, and calculate the standard deviation of these rate values. If is very small (such as less than 0.05 mm / s, indicating that the rate is very stable and judged as "high" stability), then takes a larger value; if is larger (such as greater than 0.2 mm / s, indicating that the rate fluctuation is "significant" and judged as "low" stability), then Take the smaller value. ,in is a small positive constant that prevents the denominator from being zero and adjusts the sensitivity. Assume that the last 60 seconds have been calculated (This volatility level is defined as "medium"), then .

[0066] Formation reaction force (pump pressure) weight coefficient Set to pump pressure measurement value The fluctuation variance in the past period of time (such as 60 seconds) Calculation process: Collect the pump pressure data for the past 60 seconds (one point every 2 seconds, a total of 30 points) and calculate its variance If the variance is small (e.g. less than , pressure fluctuation "low"), then Take the larger value; if the variance is large (such as greater than , pressure fluctuation "high"), then Take the smaller value. ,in For adjustment. Assuming that the calculation is (This volatility level is defined as "moderately low"), then The logic behind setting these two weight coefficients is that the higher the stability of the parameter (i.e. the smaller the rate of change or variance), the higher its representativeness and confidence in the current construction status, and therefore a greater weight is given to it. These stability ranges of "high", "medium", and "low" are determined based on statistical analysis of data from a large number of historical high-quality construction cases (selecting 100 completed and high-quality pile foundation projects). For example, statistics on the steady-state pile sinking stage of these 100 projects The distribution is defined as “high” stability if the distribution is below the 25th percentile and “low” stability if the distribution is above the 75th percentile.

[0067] Then, the system uses a dynamic weighted fusion algorithm to perform normalized linear product superposition on the two core parameter indicators to generate a comprehensive construction feedback deviation indicator. First, the current pile sinking rate and pump pressure Compare with their respective expected values or ideal intervals to obtain the normalized deviation index. is the normalized pile driving rate compliance (e.g. ,in is the target pile sinking rate at the current depth, such as ,but ), is the consistency index between the normalized pump pressure and the pressure predicted by the formation model. For simplicity, a comprehensive scoring method is adopted here. ,in and are the values of the pile driving rate and the pump pressure respectively after being standardized (such as converted into the performance scores within the range of 0 - 1, with 1 being the best). Let (indicating that the current rate shows good performance), (indicating that the current pressure state is acceptable). Then . This fusion value (in the range of 0 - 1) is compared with a preset ideal state reference value (such as , and this reference value represents the expected excellent construction state). The difference is the generated construction feedback deviation index. When the absolute value of this deviation index exceeds a preset action threshold , the system will trigger a twin correction instruction. is set to 0.15. The basis for setting this threshold: By simulating and analyzing the potential impact of different magnitudes of on the final pile foundation quality (such as bearing capacity, uniformity), when continuously exceeds 0.15 for a certain period of time (such as continuously for 5 minutes), it is considered that the construction state deviates significantly from the ideal, which may lead to an increase in quality risks. It is necessary to adjust the parameters of the digital twin model to better reflect the actual working conditions and may prompt the operator to pay attention. Currently , it does not exceed 0.15, so the correction is not triggered temporarily. If the subsequent calculated is greater than 0.15, it will be triggered.

[0068] The subgrade reaction coefficient of the Winkler foundation model is determined by inverting the data of the in-situ static cone penetration test.

[0069] Please refer to Figure 5 , the twin synchronization module includes: The thickness parameter adjustment sub-module calls the twin correction instruction, extracts the initial thickness parameters of the virtual pile body segmented model, and generates thickness correction parameters according to the inverse proportional relationship between the thickness adjustment coefficient of the correction instruction and the pile driving rate index; The thickness parameter adjustment sub-module is activated after receiving the twin correction instruction triggered by the "dynamic fusion feedback sub-module". This instruction clarifies the significant deviation type and approximate degree between the current construction state and the model prediction. Suppose the received instruction indicates: "The current actual average pile driving rate (such as ) is significantly lower than the theoretical average pile driving rate predicted by the twin model based on the current parameters (such as ). It is inferred that the reason is that the model fails to fully reflect the actual encountered formation resistance, and it is necessary to adjust the geometric parameters related to the pile - soil interaction in the model to match the actual situation." The system first extracts the "thickness parameter" of the current virtual pile segment model (P001 pile) at the previous iteration cycle or initial setting. The "thickness parameter" here is a key geometric dimension that affects the interaction area between the pile and the surrounding soil, and thus affects the calculation of side friction resistance and end bearing resistance. For a circular pile like P001, it is its equivalent action diameter . Let the current average equivalent diameter of each segment in the model be (from paragraph 7 ).

[0070] Next, the system generates a specific thickness correction parameter according to the adjustment direction indicated in the correction instruction (the resistance that needs to be increased in the model) and the actual pile sinking rate and the relationship between the theoretical pile sinking rate predicted by the model . The relationship logic between the thickness adjustment coefficient and the pile sinking rate index is as follows: When the actual pile sinking rate is significantly lower than the theoretical predicted rate, it indicates that the actual formation resistance encountered is greater than the resistance calculated by the current parameters of the model. To enable the model to reproduce this greater resistance (and thus cause the theoretical rate to decrease to approach the actual rate), it is necessary to appropriately increase the equivalent action diameter of the pile in the model (i.e., the "thickness parameter"), because an increase in diameter will lead to an increase in the side friction resistance calculation area and the end bearing area (or its influence coefficient). The thickness adjustment amount is calculated as follows: . Among them, is a calibrated adjustment sensitivity coefficient, whose value is determined based on a large number of simulation tests and back-analysis of historical data, and is used to control the amplitude of a single adjustment to avoid model oscillation is set to 0.15. The basis for setting this value: Through retrospective analysis of 20 historical pile foundation cases, it is found that when using the value within the range of 0.1 to 0.2 for iterative correction in similar deviation situations, the model convergence speed and stability are better. Calculate: ; This calculated is the thickness (diameter) correction increment for this iteration. Therefore, the new equivalent action diameter of P001 pile in the twin model will be adjusted to . The purpose of this adjustment is to enable the digital twin model to more accurately simulate the actually observed pile sinking behavior (i.e., a lower pile sinking rate) in the next iteration calculation

[0071] The settlement rate update sub-module calculates the absolute value of the difference between the actual settlement rate and the theoretical rate based on the thickness correction parameter, and uses cubic spline interpolation to distribute the difference to the corresponding model nodes according to the time section length to generate a dynamic thickness update value The settlement rate update sub-module is based on the thickness correction parameter generated by the "thickness parameter adjustment sub-module" in the previous step and the adjusted new equivalent acting diameter to perform operations. The system also records the theoretical average settlement rate predicted by the model based on the old diameter before applying this thickness correction , as well as the average settlement rate actually observed during the same evaluation period .

[0072] The system first uses the updated equivalent acting diameter of the pile , combined with the existing formation parameters and load conditions, to perform a new complete pile driving process simulation analysis or simplified theoretical calculation in the digital twin model, so as to obtain a new theoretical average settlement rate corresponding to the new diameter . Since the pile diameter increases, the total formation resistance (mainly the side friction resistance) calculated by the model will increase accordingly, so the new theoretical settlement rate is expected to decrease. After calculation, the new theoretical average settlement rate .

[0073] Then, the system calculates the absolute difference between this new theoretical settlement rate and the average settlement rate actually observed . . This residual difference represents the difference that still exists between the settlement rate predicted by the model and the actual observed rate after adjusting the model parameters (equivalent diameter) once. Ideally, this difference should approach zero

[0074] Next, the system uses the cubic spline interpolation method to smoothly and reasonably distribute the change in the pile geometry caused by the adjustment of the overall equivalent diameter from to (total increment ) to each segmented model node along the entire length (30.15 m) of the virtual pile P001, forming specific dynamic thickness (diameter) update values for each segment. The use of cubic spline interpolation aims to ensure the smoothness and continuity of the change in the pile diameter along the depth, and avoid unnatural geometric mutations at the segment joints. The specific operation is as follows: Divide the P001 pile along the depth direction into A control node (for example, one control node is set every 1 meter, with a total of about 30 nodes). The diameter values before adjustment at each control node are known, as well as the overall average diameter increase to be achieved. The cubic spline interpolation algorithm will construct a series of piecewise cubic polynomial functions. These functions are not only continuous in terms of function values (diameter) at each control node, but also their first derivatives (diameter change rate) and second derivatives (change rate of diameter change rate) are continuous. By solving the coefficients of these polynomials, the specific increment of the diameter at each control node can be obtained These weighted average effects (taking into account the lengths of each segment) should be able to achieve the goal of increasing the overall equivalent diameter by approximately 45.17 mm. For example, for the segments in the middle region of the pile body, the assigned diameter increment may be slightly higher than the average value, while the increments in the regions near the pile top and pile bottom may be slightly lower, depending on the boundary condition settings of the interpolation algorithm and the preset deformation mode (such as assuming that the pile body deformation mainly occurs in the middle). Finally, the system generates an accurate sequence of dynamic thickness (diameter) update values for each segment or key node in the P001 pile model. These update values will be directly used to modify the geometric parameters of the twin model so that it can better match the actual working conditions in the next simulation or analysis

[0075] The curve shape matching sub-module obtains the time-domain sampling point sequences of the injection rate curve and the reaction force measurement curve, calculates the sum of the squares of the slope differences between the two curves at multiple time nodes, filters the nodes that exceed the shape similarity threshold, reallocates the weight ratio of the dynamic thickness update values, and generates updated twin parameters

[0076] The curve shape matching sub-module obtains the time-domain sampling point sequences of two types of key dynamic parameters during the construction process: one is the injection rate curve. For the static pressure pile driving process of P001 pile, this refers to the actual penetration depth change rate of the pile body per unit time, that is, the real-time pile driving rate calculated above ; the other is the reaction force measurement curve, which refers to the total jacking pressure monitored in real time by the hydraulic system of the static pressure pile driver , and this pressure directly reflects the total formation reaction force received by the pile body during the sinking process. The system extracts these two groups of accurately time-aligned data sequences from the construction monitoring database, and the sampling time interval is 2 seconds. Injection rate (pile driving rate) sequence: . Reaction force measurement (total jacking pressure) sequence: . (equivalent to 12.3 MPa pump pressure) ; The system calculates these two time sequences at each sampling time point The instantaneous change trend at a certain point, that is, calculate its first-order difference or local slope. The change amount of the pile driving rate: ; The change amount of the total jacking pressure: ; At this point (using the subsequent point to calculate the change): (The rate is increasing) (The pressure is increasing); Next, the system evaluates the morphological similarity of these two sets of change amount sequences and . A key evaluation index is the consistency of their change directions. Under normal circumstances, when the pile driving rate increases (the pile penetration accelerates), if the formation conditions remain unchanged, the jacking pressure may decrease slightly or remain stable (inertia effect or soil disturbance softening); but more commonly, if the rate is to be maintained or increased, especially in relatively hard soil layers, the jacking pressure needs to increase accordingly. Conversely, if the jacking pressure increases significantly, the pile driving rate usually slows down. Here, a morphological similarity evaluation rule is set: within a certain time window (such as 5 consecutive sampling points, that is, 10 seconds), if and 's signs (positive and negative signs, representing the increasing and decreasing trends) show an "opposite" relationship at more than 60% of the points (that is, at least 3 points) (for example, the rate increases while the pressure decreases, or the rate decreases while the pressure increases, which may indicate special soil responses or construction control adjustments), or the amplitude ratio of their co-directional changes ( ) significantly deviates from the normal range predicted by the current formation model (this range is obtained through model simulation, for example, the normal range is from 10 kN / (mm / s) to 30 kN / (mm / s)), then the node morphological matching degree within this time window is considered "low". The morphological similarity threshold is defined as: within a 10-second window, if the number of "opposite" trend points exceeds 3, or 's ratio exceeds the interval of [5, 35] kN / (mm / s) (this interval is statistically based on historical data and represents 80% of normal construction conditions), then it is determined as morphological mismatch.

[0077] The system screens out all time nodes or sections determined to be morphologically mismatched. For these sections with "low" morphological matching degrees, the system will re-examine and adjust the allocation weight ratio of the dynamic thickness (diameter) update value generated by the "settlement rate update sub-module". The specific operation is: in those sections where the response patterns of the pile driving rate and the jacking pressure show inconsistencies with the predicted behavior of the normal model (such as period, the rate suddenly increases greatly while the pressure does not increase accordingly, which may imply a sudden drop in local soil bearing capacity or special situations in the hole), the system will reduce the application weight of the thickness (diameter) correction amount of this section previously calculated mainly based on the overall settlement rate deviation (for example, the originally calculated The weight is adjusted from 1.0 to 0.4), and an additional, targeted local model parameter adjustment factor (such as local soil parameters or pile-soil contact model parameters) may be introduced according to the characteristics of the current morphological mismatch (such as pressure response lag or over-sensitivity), rather than just adjusting the diameter. Through this weight reallocation and refined adjustment based on curve morphology matching for the dynamically updated thickness value (and possibly other associated model parameters), a comprehensive parameter set for updating the digital twin model is finally generated. This ensures that the updated twin not only matches the actual situation in terms of the overall average settlement behavior, but also can more realistically reproduce the real-time construction conditions in terms of the key rate-reaction dynamic response characteristics reflecting complex pile-soil interactions.

[0078] The above is only a preferred embodiment of the present invention, and does not limit the present invention in other forms. Any person skilled in the art may use the technical content disclosed above to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, as long as it does not depart from the technical solution content of the present invention, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. A digital twin system for the whole process of pile foundation construction, characterized in that The system includes: A twin configuration module, which is used to generate a virtual pile segment model through the burial depth coordinate sequence and the concrete pouring timestamp, call the spatio-temporal interpolation algorithm to construct a time-axis driven model sequence, bind the steel cage segment number and the timestamp, establish a dynamic mapping relationship, generate a virtual pile segment model and transfer it to the dynamic mapping module; A dynamic mapping module, which is used to extract the axis deviation value of the virtual pile segment model, calculate the Euclidean distance by unifying the coordinate unit, call the verticality threshold to judge the abnormal nodes, input the point cloud registration algorithm to correct the coordinates, generate a dynamic topological twin and transfer it to the coupling analysis module; A coupling analysis module, which is used to receive the dynamic topological twin, solve the theoretical value of the formation reaction force and calculate the pile sinking rate index, dynamically weighted fuse the injection rate and the measured value of the formation reaction force to generate the perfusion coupling parameter, execute the dynamic weighted fusion algorithm to generate the construction feedback deviation index, trigger the twin correction instruction and transfer it to the twin synchronization module.

2. The digital twin system for the whole process of pile foundation construction according to claim 1, wherein The spatio-temporal interpolation algorithm synchronously updates the timestamp label of the model attribute node with the construction pouring progress by performing hash matching on the steel cage segment number and the timestamp; The verticality threshold is 5% of the pile diameter and is determined by verifying the on-site test pile data; The dynamic weighted fusion algorithm makes the error between the construction feedback deviation index and the actual deformation of the pile body less than 3% by normalizing the linear product and superimposing the two types of parameters.

3. The digital twin system for the whole process of pile foundation construction according to claim 2, characterized in that, The twin configuration module includes: The coordinate time series processing sub-module obtains the burial depth coordinate sequence and the concrete pouring timestamp, performs spatial hierarchical matching on the longitude, latitude, and elevation components of the coordinate points and the timestamp, calculates the difference between adjacent timestamp intervals based on the moving speed of the pouring machinery, and integrates the spatial nodes and the difference parameters to generate a pile segment coordinate set; The spatio-temporal interpolation modeling sub-module divides the three-dimensional grid unit based on the node distribution density of the pile segment coordinate set, uses the Lagrange interpolation method to construct a time-axis increment sequence for the uncovered area, corrects the continuity of the grid vertex coordinates according to the timestamp label bound to the steel cage, and outputs a time-axis driven parameter group; The steel cage time binding sub-module extracts the steel cage segment number from the time-axis driven parameter group, maps the segment number to the segmented interval corresponding to the timestamp, and uses the hash matching algorithm to associate the segment number with the timestamp label of the model attribute node to generate a virtual pile segment model; The spatial hierarchical matching is realized by establishing a real-time association between the longitude and latitude coordinates and the GNSS positioning data of the construction machinery.

4. The digital twin system for the whole process of pile foundation construction according to claim 3, characterized in that, The dynamic mapping module includes: The axis deviation extraction sub-module obtains the axis coordinate sequence of the virtual pile segment model, decomposes the longitude, latitude, and elevation components of the model nodes, extracts the corresponding coordinate values of the theoretical axis, and uses the Euclidean distance formula to calculate the absolute value of the difference between the actual coordinate and the theoretical coordinate to generate an axis coordinate deviation value; The verticality verification sub-module, based on the axis coordinate deviation value, calls the axial verticality threshold, compares the deviation value with the threshold item by item in the node order, marks the spatial index of the node where the deviation value exceeds the axial verticality threshold, and generates an abnormal node identification set; The point cloud correction sub-module calls the point cloud registration algorithm to process the abnormal node identification set, extracts the coordinate offsets of the abnormal nodes, matches the spatial distribution of the abnormal points and the theoretical point cloud through the iterative closest point algorithm, and introduces the pile elastic deformation parameter to constrain the coordinate correction range, generating a dynamic topological twin body.

5. The digital twin system for the whole process of pile foundation construction according to claim 4, wherein The theoretical axis coordinate value is obtained by calculating the average value of the pile center coordinate on the design drawing and the construction lofting point coordinate.

6. The digital twin system for the whole process of pile foundation construction according to claim 5, characterized in that, The coupling analysis module includes: The formation reaction solution operator module receives the dynamic topological twin body, decomposes the pile diameter, pile length, and material elastic modulus in the twin body structure parameters, applies the Winkler foundation model to establish a linear equation between the formation reaction and the pile displacement, calculates the superposition theoretical value of the pile tip resistance and the side friction resistance, and generates the formation reaction theoretical value. The pile driving rate calculation sub-module calls the real-time pump pressure sensor data and the settlement amount measured by the laser rangefinder, converts the pump pressure value into the pile tip pressure component in kilonewtons and unifies it to the second-level time reference with the settlement amount, constructs an equilibrium equation between the pile driving resistance and the pump pressure based on the formation reaction theoretical value, and calculates the settlement amount increment per unit time using the forward difference method, outputting the pile driving rate index. The dynamic fusion feedback sub-module sets the injection rate weight coefficient as the reciprocal of the pump pressure change rate and the formation reaction weight coefficient as the reciprocal of the measured value fluctuation variance based on the pile driving rate index and the pressure sensor measured value, performs normalized linear product superposition on the two types of parameters through the dynamic weighted fusion algorithm, generates the construction feedback deviation index, and triggers the twin body correction instruction.

7. The digital twin system for the whole process of pile foundation construction according to claim 6, characterized in that, The bed coefficient of the Winkler foundation model is determined by inverting the field static cone penetration test data.

8. The digital twin system for the whole process of pile foundation construction according to claim 7, characterized in that, The system further includes: The twin synchronization module is used to adjust the thickness parameter of the virtual pile body segmented model based on the twin body correction instruction, update the thickness according to the settlement rate difference, match the similarity of the curve shapes of the injection rate and the reaction measured value, generate the updated twin body parameters, and synchronize them backward to the dynamic mapping module and the coupling analysis module.

9. The digital twin system for the whole process of pile foundation construction according to claim 8, wherein, The updated twin body parameters specifically refer to the adjusted pile segment thickness parameter and the material layer parameter. The curve shape similarity is calculated by calculating the sum of the squares of the slope differences of the two curves at adjacent time nodes, and screening the nodes with differences exceeding 0.5 for weight ratio adjustment.

10. The digital twin system for the whole process of pile foundation construction according to claim 9, characterized in that, The twin synchronization module includes: The thickness parameter adjustment sub-module calls the twin body correction instruction, extracts the initial thickness parameter of the virtual pile body segmented model, and generates the thickness correction parameter according to the inverse proportional relationship between the thickness adjustment coefficient of the correction instruction and the pile driving rate index. The settlement rate update sub-module calculates the absolute value of the difference between the actual settlement rate and the theoretical rate based on the thickness correction parameter, and uses the cubic spline interpolation method to distribute the difference to the corresponding model nodes according to the time section length, generating the dynamic thickness update value. The curve shape matching sub-module obtains the time-domain sampling point sequences of the injection rate curve and the reaction measured value curve, calculates the sum of the squares of the slope differences of the two curves at multiple time nodes, screens the nodes exceeding the curve shape similarity threshold, and redistributes the weight ratio of the dynamic thickness update value to generate the updated twin body parameters.

Images