Online finite element bridge monitoring modeling method and system

Abstract

The invention provides an online finite element bridge monitoring modeling method and system, and the method comprises the steps: applying the same load condition to a bridge three-dimensional model, and obtaining the simulation displacement data of the bridge three-dimensional model through a displacement change algorithm based on a monitoring threshold value; comparing the monitored displacement data with the simulated displacement data to obtain an analysis error; and adjusting a monitoring threshold value of the bridge three-dimensional model based on the analysis error. According to the method, conversion from the MidasMCT finite element data file to the OpenSees TCL finite element data file can be processed online through the Internet, and conversion of nodes, sections and linearity of the bridge structure is achieved. Real-time monitoring displacement data accessed to bridge monitoring can be compared with simulation data of the finite element, and the accuracy of a finite element simulation model is further verified, so that a bridge monitoring threshold value is adjusted.

Description

Technical Field

[0001] This invention belongs to the field of engineering monitoring technology, specifically relating to an online finite element bridge monitoring modeling method and system. Background Technology

[0002] Currently, in the field of bridge monitoring, commercial finite element software such as Midas, Abaqus, Ansys, and BridgeDoctor are commonly used for structural modeling and threshold calculation. However, these software programs are often available in desktop form and cannot provide real-time data interfaces to communicate with cloud platforms such as bridge monitoring platforms. Therefore, the models and thresholds can only be re-entered into the bridge monitoring cloud platform and cannot be dynamically updated. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention provides an online finite element bridge monitoring and modeling method and system, which can effectively solve the above-mentioned problems.

[0004] The technical solution adopted in this invention is as follows:

[0005] This invention provides an online finite element bridge monitoring and modeling method, comprising:

[0006] Step S1: Convert the bridge MidasMCT finite metadata file into an OpenSeesTCL finite metadata file;

[0007] Step S2: The browser loads the OpenSeesTCL finite metadata file and parses the OpenSeesTCL finite metadata file into a visualized 3D model of the bridge.

[0008] Step S3: Real-time access to the load conditions applied to the bridge and the monitoring displacement data caused by the applied load conditions through the bridge monitoring data interface.

[0009] Step S4: Apply the same load condition to the three-dimensional bridge model, and obtain the simulated displacement data of the three-dimensional bridge model through a displacement change algorithm based on the monitoring threshold; compare the monitored displacement data and the simulated displacement data to obtain the analysis error; and adjust the monitoring threshold of the three-dimensional bridge model based on the analysis error.

[0010] Furthermore, step S1 specifically includes:

[0011] Step S11: Establish a unified data structure;

[0012] Step S12: Perform multimodal parsing and semantic understanding on the bridge's MidasMCT finite metadata file. Through topology reconstruction, convert each bridge component along its axis into node-element units. Further analysis yields the following attributes for each element: element size unit, element force unit, element material, element cross-section, node number within the element, constraints of the element, loads applied to the element, and connection method between elements. Store all attributes of each parsed element into the unified data structure. After all elements have been parsed, a structured dictionary is obtained, forming a structured middleware platform.

[0013] Step S13: Perform unit consistency and dimension mapping transformation on the data stored in the structured dictionary to obtain the transformed structured dictionary;

[0014] Step S14: Based on the transformed structured dictionary, perform topology reconstruction transformation and local coordinate reconstruction of each element, perform consistent modeling and expression of the cross section and modeling and expression of material properties, and obtain the OpenSeesTCL finite metadata file.

[0015] Furthermore, the multimodal parsing of the bridge's MidasMCT finite metadata file includes:

[0016] Standardized mapping and automatic modeling of materials and cross sections: On the material side, material properties are distinguished and unit-consistent conversion and elastic modulus mapping are performed; the material properties include concrete and steel materials; on the cross section side, database cross sections, variable cross sections and numerical cross sections are identified and classified, and the cross sections include outer contours and multiple internal hole point sets; the identified cross sections are automatically generated to generate fiber cross sections and end-consistent modeling.

[0017] The connection side analyzes rigid and elastic connection methods, and generates material parameters and coefficients according to the degree of freedom of the connection method.

[0018] The static load and moving load are discretized into a node-step scheduling table on the load side.

[0019] Furthermore, the data stored in the structured dictionary undergoes unit consistency and dimensional mapping transformation, specifically as follows:

[0020] Force unit conversion factor Conversion factor for length units The following formula is used for unit unification and dimensional mapping transformation:

[0021]

[0022] in: , , , , , These are the converted force, length, area, moment of inertia, elastic modulus, and elastic coefficient, respectively. , , , , , These represent the force, length, area, moment of inertia, elastic modulus, and elastic coefficient before conversion, respectively; the unit for force after conversion is Newtons (N); and the unit for length after conversion is millimeters (mm).

[0023] Furthermore, the topology reconstruction transformation and the local coordinate reconstruction of each cell are specifically as follows:

[0024] Establish a global coordinate system; transform each type of bridge component into a node-element along its axis, and create a three-dimensional coordinate table of nodes in the global coordinate system; establish the connectivity between nodes and elements to form a unified topology;

[0025] For each element in the global coordinate system, a local coordinate system for each element is established. The method for establishing the local coordinate system is as follows:

[0026] The three-dimensional global coordinates of the two nodes connected by the unit are: First node Second node Then define the local x-axis of the local coordinate system as x_raw = x = x_raw / ||x_raw||;

[0027] Specify a reference vector v_raw that is not collinear with the local x-axis x_raw, and calculate the local z-axis z_raw = x_raw × v_raw; z = z_raw / ||z_raw||.

[0028] This leads to the establishment of a local coordinate system for the element.

[0029] Furthermore, the consistent modeling and representation of the cross-section specifically includes:

[0030] Analyze the cross-sectional parameters to extract the cross-sectional type and shape parameters;

[0031] Based on the cross-section type and cross-section shape parameters, regular mapping and classification processing are performed:

[0032] ①The cross-section type is a numerical cross-section:

[0033] If the cross-section type is a numerical cross-section, calculate the outer contour area A_outer and the area of ​​each inner hole; subtract the outer contour area A_outer from the area of ​​all inner holes to obtain the effective cross-section area A_eff;

[0034] Use the ray casting method to determine the number of valid sampling points N_valid on the outer contour of the cross section that are not in any inner hole;

[0035] The equivalent area of ​​a single fiber is obtained using the formula A_fiber = A_eff / N_valid;

[0036] For each valid sampling point, generate fibers of equal area, with the material number inherited from the parent material;

[0037] This allows the numerical cross-section with multiple cavities to be stably mapped to a fiber cross-section, maintaining area consistency.

[0038] ②The cross-section type is variable cross-section:

[0039] The method for consistent modeling of variable cross sections is as follows:

[0040] The local coordinates of the principal axes of the beam element are ξ∈[0, 1]. For the starting section P0 and the ending section P1 at both ends of the beam, linear equivalent interpolation is performed on the points of the beam element along the beam length to construct a variable cross section with consistent modeling expression at the starting and ending ends.

[0041] Therefore, for cross-sections with different geometries, the material-section association lookup module binds the material number to the cross-section parameters, calls the corresponding cross-section generation module to construct the fiber cross-section, and automatically outputs the cross-section definition as the structural analysis script. For numerical cross-sections and variable cross-sections of prestressed concrete, fiber subdivision is constructed based on the outer contour points and the set of multiple inner hole contours, supporting consistent modeling expression of the start and end ends. At the same time, anomaly alarms and degradation output strategies are provided to ensure that compatible cross-section definitions can still be generated when parameters are incomplete or parsing fails.

[0042] Specifically, it includes a data parsing and parameter extraction module, a material mapping module, a shape recognition and strategy distribution module, a fiber mesh generation module, a script writing module, and an exception rollback module, forming an scalable shape plugin mechanism and a robust engineering implementation process.

[0043] Furthermore, the modeling and expression of material properties specifically includes:

[0044] From the transformed structured dictionary, the material definition is read, and the material definition is normalized and mapped to the material constitutive model that can be executed by the analysis solver. The specific process is as follows: read the material number and the set of material properties, classify and process them according to the material type and convert them to unit consistency, and distinguish between rigid connection and general elastic connection for connection materials.

[0045] Output the elastic material definition containing Poisson's ratio for user-defined materials, read the elastic modulus and Poisson's ratio, and unify them to N / mm²;

[0046] For concrete and steel, the standard elastic modulus is used, and the grade is mapped to the target elastic modulus.

[0047] If the connecting materials are general elastic connections, determine the start node and end node in the unit formed by the general elastic connection, the material labels corresponding to each of the 6 degrees of freedom, generate an independent uniaxial material in each degree of freedom direction, and use the material with E=0 if the stiffness is 0.

[0048] If the connecting materials are rigid, the large E approximation is used. In the element formed by the connecting materials, the master node and slave node are determined. Through multi-point constraint commands, the slave node and the master node are forced to maintain the same motion in the six degrees of freedom that need to be constrained, thus obtaining the modeled element of the connecting materials. The six degrees of freedom include X translation, Y translation, Z translation, rotation about X, rotation about Y, and rotation about Z.

[0049] Furthermore, in step S4, the same load condition is applied to the three-dimensional bridge model, and simulated displacement data of the three-dimensional bridge model is obtained through a displacement change algorithm based on a monitoring threshold, specifically as follows:

[0050] It reads static and moving load information, automatically converts it into a recognizable command stream, and achieves integrated, online modeling and solving of static and moving loads by programmatically editing and rewriting the load mode segments in the input script.

[0051] For static loads, automatically generate or update load patterns and nodal load lists, apply the concentrated force / moment of a node to the structure, and obtain the equilibrium displacement and internal forces through static step calculation. Data items: node number, force component and moment component.

[0052] For moving loads, the system discretizes vehicle axle load, speed, and travel path into time steps or position steps, constructs a load scheduling table that migrates on the node sequence over time or steps, and performs the solution step by step. When the user modifies the load, the system only rewrites the load-related segments and triggers incremental solution, achieving efficient, rollbackable, and observable moving load analysis.

[0053] The moving load includes a basic parameter configuration module, a load editor, a moving load scheduler algorithm, a solver orchestrator, and a result streaming service. The load editor is used for incremental script modification and supports default rollback and versioning. The solver orchestrator has modules for static stepping, error monitoring, failure retry, and rollback, updating the model state only when necessary. The result streaming service pushes displacement and internal force data to the front end in real time.

[0054] Furthermore, the specific implementation method for the moving load is as follows:

[0055] 1) Parameter configuration:

[0056] A. The lane path is discretized into nodes, forming a path node set pathNodes=[n0, n1, ..., n K ]; where n0, n1, ..., n K These are the nodes arranged in order;

[0057] B. Construct the vehicle-axle group parameter set axles=[{axleId,weight_N,axleSpacing,laneOffset,timeOffset}];

[0058] Where: axleId represents the axle number; weight_N represents the axle weight; axleSpacing represents the wheelbase; laneOffset represents the lane offset; and timeOffset represents the time offset.

[0059] C. Global timing parameters: step size dt, vehicle speed v, and node spacing dx;

[0060] D. The loading list for each analysis step k is Schedule[k] = [{nodeId, Fz, Fx, Fy, Mx, My, Mz,laneId, axleId}];

[0061] Where: nodeId is the node number, Fz is the force in the Z direction, Fx is the force in the X direction, Fy is the force in the Y direction, Mx is the torque about the X axis, My is the torque about the Y axis, Mz is the torque about the Z axis, laneId is the lane number, and axleId is the axle number.

[0062] E. Generation principle: Each node corresponds to a valid node length;

[0063] For each lane and each axle, calculate the length of the effective node to which it belongs in step k, and then determine the node to which it belongs. Write the relevant parameters into the schedule[k] for this step; multiple lanes / multiple axles are naturally superimposed in the same step k.

[0064] 2) Implementation of the moving load scheduler algorithm:

[0065] A. Dynamic position calculation algorithm for multi-axle vehicle loads:

[0066] a. During vehicle movement, based on the vehicle speed v and the cumulative wheelbase offset offset_j of each axle j relative to the front axle, the position of each axle j at step k is obtained. ;

[0067] b. Determine the node to which each axle j belongs based on its position x_axle(k) at step k;

[0068] c. Within the range of axle j from the initial step k_start of the effective node length entering node to the final step k_end of the effective node length leaving node, the total cumulative load F_total of axle j acting on node is obtained using the continuous-time integral formula for moving load:

[0069] F_total=

[0070] in: Let be the load value of axle j at step k, which is a constant;

[0071] Let be a position function, representing the load value of axle j at step k in spatial position. The centralizing effect of the location;

[0072] exist When the value is infinity, and at other locations it is 0. The integral represents the load acting at the location. ;

[0073] b. Dynamic response of the structure under moving loads:

[0074] u(k) =

[0075] The meaning is: the dynamic response u(k) of each step k, that is, the displacement of the structure at step k due to the moving load, is the sum of the contributions of all historical loads in the current step k within the time range from 0 to step k.

[0076] Each step within the time range of 0 to step k ,step load The contribution of step k is determined by the load. The size of the distance and the time interval of step k are determined by the magnitude of the distance.

[0077] Let be the unit impulse response function, representing the response at step . The response generated at step k when a unit pulse load is applied.

[0078] This invention also provides a system for implementing the aforementioned online finite element bridge monitoring and modeling method, comprising:

[0079] The conversion module is used to convert the bridge MidasMCT finite metadata file into an OpenSeesTCL finite metadata file;

[0080] The parsing module is used to enable the browser to load the OpenSeesTCL finite metadata file and parse the OpenSeesTCL finite metadata file into a visualized 3D model of the bridge.

[0081] The load loading and displacement monitoring module is used to access the load conditions applied to the bridge and the monitoring displacement data caused by the bridge under the applied load conditions in real time through the bridge monitoring data interface.

[0082] The bridge 3D model simulation module is used to apply the same load condition to the bridge 3D model, obtain the simulated displacement data of the bridge 3D model through a displacement change algorithm based on a monitoring threshold, compare the monitored displacement data and the simulated displacement data to obtain the analysis error, and adjust the monitoring threshold of the bridge 3D model based on the analysis error.

[0083] The online finite element bridge monitoring and modeling method and system provided by this invention has the following advantages:

[0084] 1. This invention proposes a finite element transformation method based on OpenSees, which can be implemented via the Internet.

[0085] The line processing converts MidasMCT finite metadata files to OpenSees TCL finite metadata files, enabling the conversion of bridge structure nodes, sections, and lines.

[0086] 2. Through the analysis of static and moving loads, the displacement calculation of key bridge components under different working conditions was realized using OpenSees TCL finite element metadata files. Compared with the commercial finite element software Midas, due to the different algorithms, the error of simply supported beam bridges was controlled within 5%, and the error of large and complex bridges was controlled within 10%.

[0087] 3. By using lightweight web technology, the limited metadata files of OpenSees TCL were loaded in the browser, enabling lightweight online access.

[0088] 4. By accessing the real-time monitoring displacement data in bridge monitoring, it is possible to compare it with the simulation data of the finite element method of this invention, further verifying the accuracy of the finite element simulation model, so as to adjust the bridge monitoring threshold. Attached Figure Description

[0089] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0090] Figure 1 This is a flowchart of an online finite element bridge monitoring and modeling method according to the present invention;

[0091] Figure 2 The diagram shows the finite element analysis results provided for an embodiment of the present invention. Detailed Implementation

[0092] To make the technical problems solved, the technical solutions, and the beneficial effects of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and are not intended to limit the invention.

[0093] This invention can import existing finite element models of bridges into a monitoring system for lightweight modeling, enabling rapid and lightweight modeling of bridge points, alignments, cross-sections, materials, static loads, and moving loads. It allows for backend calculations and web visualization, offering advantages such as online operation, real-time performance, rapid processing, and strong scalability. Furthermore, this invention can convert commonly used Midas MCT commercial finite element metadata files to OpenSees TCL finite element metadata files, and achieve dynamic 3D finite element modeling and dynamic threshold calculation.

[0094] See Figure 1 This invention provides an online finite element bridge monitoring and modeling method, comprising:

[0095] Step S1: Convert the bridge MidasMCT finite metadata file into an OpenSeesTCL finite metadata file;

[0096] Step S2: The browser loads the OpenSeesTCL finite metadata file and parses the OpenSeesTCL finite metadata file into a visualized 3D model of the bridge.

[0097] Step S3: Real-time access to the load conditions applied to the bridge and the monitoring displacement data caused by the applied load conditions through the bridge monitoring data interface.

[0098] Step S4: Apply the same load condition to the three-dimensional bridge model, and obtain the simulated displacement data of the three-dimensional bridge model through a displacement change algorithm based on the monitoring threshold; compare the monitored displacement data and the simulated displacement data to obtain the analysis error; and adjust the monitoring threshold of the three-dimensional bridge model based on the analysis error.

[0099] The following is a detailed description of each step:

[0100] Step S1: Convert the bridge MidasMCT finite metadata file into an OpenSeesTCL finite metadata file;

[0101] Step S1 is as follows:

[0102] Step S11: Establish a unified data structure;

[0103] Step S12: Perform multimodal parsing and semantic understanding on the bridge's MidasMCT finite metadata file. Through topology reconstruction, convert each bridge component along its axis into node-element units. Further analysis yields the following attributes for each element: element size unit, element force unit, element material, element cross-section, node number within the element, constraints of the element, loads applied to the element, and connection method between elements. Store all attributes of each parsed element into the unified data structure. After all elements have been parsed, a structured dictionary is obtained, forming a structured middleware platform.

[0104] In this step, the multimodal parsing of the bridge's MidasMCT finite metadata file includes:

[0105] Standardized mapping and automatic modeling of materials and cross sections: On the material side, material properties are distinguished and unit-consistent conversion and elastic modulus mapping are performed; the material properties include concrete and steel materials; on the cross section side, database cross sections, variable cross sections and numerical cross sections are identified and classified, and the cross sections include outer contours and multiple internal hole point sets; the identified cross sections are automatically generated to generate fiber cross sections and end-consistent modeling.

[0106] The connection side analyzes rigid and elastic connection methods, and generates material parameters and coefficients according to the degree of freedom of the connection method.

[0107] The static load and moving load are discretized into a node-step scheduling table on the load side.

[0108] Step S13: Perform unit consistency and dimension mapping transformation on the data stored in the structured dictionary to obtain the transformed structured dictionary;

[0109] This step is specifically as follows:

[0110] Force unit conversion factor Conversion factor for length units The following formula is used for unit unification and dimensional mapping transformation:

[0111]

[0112] in: , , , , , These are the converted force, length, area, moment of inertia, elastic modulus, and elastic coefficient, respectively. , , , , , These represent the force, length, area, moment of inertia, elastic modulus, and elastic coefficient before conversion, respectively; the unit for force after conversion is Newtons (N); and the unit for length after conversion is millimeters (mm).

[0113] Step S14: Based on the transformed structured dictionary, perform topology reconstruction transformation and local coordinate reconstruction of each element, perform consistent modeling and expression of the cross section and modeling and expression of material properties, and obtain the OpenSeesTCL finite metadata file.

[0114] The topology reconstruction transformation and the local coordinate reconstruction of each cell are specifically as follows:

[0115] Establish a global coordinate system; transform each type of bridge component into a node-element along its axis, and create a three-dimensional coordinate table of nodes in the global coordinate system; establish the connectivity between nodes and elements to form a unified topology;

[0116] For each element in the global coordinate system, a local coordinate system for each element is established. The method for establishing the local coordinate system is as follows:

[0117] The three-dimensional global coordinates of the two nodes connected by the unit are: First node Second node Then define the local x-axis of the local coordinate system as x_raw = x = x_raw / ||x_raw||;

[0118] Specify a reference vector v_raw that is not collinear with the local x-axis x_raw, and calculate the local z-axis z_raw = x_raw × v_raw; z = z_raw / ||z_raw||.

[0119] This leads to the establishment of a local coordinate system for the element.

[0120] The consistent modeling and representation of the cross-section specifically includes:

[0121] Analyze the cross-sectional parameters to extract the cross-sectional type and shape parameters;

[0122] Based on the cross-section type and cross-section shape parameters, regular mapping and classification processing are performed:

[0123] ①The cross-section type is a numerical cross-section:

[0124] If the cross-section type is a numerical cross-section, calculate the outer contour area A_outer and the area of ​​each inner hole; subtract the outer contour area A_outer from the area of ​​all inner holes to obtain the effective cross-section area A_eff;

[0125] Use the ray casting method to determine the number of valid sampling points N_valid on the outer contour of the cross section that are not in any inner hole;

[0126] The equivalent area of ​​a single fiber is obtained using the formula A_fiber = A_eff / N_valid;

[0127] For each valid sampling point, generate fibers of equal area, with the material number inherited from the parent material;

[0128] This allows the numerical cross-section with multiple cavities to be stably mapped to a fiber cross-section, maintaining area consistency.

[0129] ②The cross-section type is variable cross-section:

[0130] The method for consistent modeling of variable cross sections is as follows:

[0131] The local coordinates of the principal axes of the beam element are ξ∈[0, 1]. For the starting section P0 and the ending section P1 at both ends of the beam, linear equivalent interpolation is performed on the points of the beam element along the beam length to construct a variable cross section with consistent modeling expression at the starting and ending ends.

[0132] Therefore, for cross-sections with different geometries, the material-section association lookup module binds the material number to the cross-section parameters, calls the corresponding cross-section generation module to construct the fiber cross-section, and automatically outputs the cross-section definition as the structural analysis script. For numerical cross-sections and variable cross-sections of prestressed concrete, fiber subdivision is constructed based on the outer contour points and the set of multiple inner hole contours, supporting consistent modeling expression of the start and end ends. At the same time, anomaly alarms and degradation output strategies are provided to ensure that compatible cross-section definitions can still be generated when parameters are incomplete or parsing fails.

[0133] Specifically, it includes a data parsing and parameter extraction module, a material mapping module, a shape recognition and strategy distribution module, a fiber mesh generation module, a script writing module, and an exception rollback module, forming an scalable shape plugin mechanism and a robust engineering implementation process.

[0134] The modeling and expression of material properties are specifically as follows:

[0135] From the transformed structured dictionary, the material definition is read, and the material definition is normalized and mapped to the material constitutive model that can be executed by the analysis solver. The specific process is as follows: read the material number and the set of material properties, classify and process them according to the material type and convert them to unit consistency, and distinguish between rigid connection and general elastic connection for connection materials.

[0136] Output the elastic material definition containing Poisson's ratio for user-defined materials, read the elastic modulus and Poisson's ratio, and unify them to N / mm²;

[0137] For concrete and steel, the standard elastic modulus is used, and the grade is mapped to the target elastic modulus.

[0138] If the connecting materials are general elastic connections, determine the start node and end node in the unit formed by the general elastic connection, the material labels corresponding to each of the 6 degrees of freedom, generate an independent uniaxial material in each degree of freedom direction, and use the material with E=0 if the stiffness is 0.

[0139] If the connecting materials are rigid, the large E approximation is used. In the element formed by the connecting materials, the master node and slave node are determined. Through multi-point constraint commands, the slave node and the master node are forced to maintain the same motion in the six degrees of freedom that need to be constrained, thus obtaining the modeled element of the connecting materials. The six degrees of freedom include X translation, Y translation, Z translation, rotation about X, rotation about Y, and rotation about Z.

[0140] Step S2: The browser loads the OpenSeesTCL finite metadata file and parses the OpenSeesTCL finite metadata file into a visualized 3D model of the bridge.

[0141] Step S3: Real-time access to the load conditions applied to the bridge and the monitoring displacement data caused by the applied load conditions through the bridge monitoring data interface.

[0142] Step S4: Apply the same load condition to the three-dimensional bridge model, and obtain the simulated displacement data of the three-dimensional bridge model through a displacement change algorithm based on the monitoring threshold; compare the monitored displacement data and the simulated displacement data to obtain the analysis error; and adjust the monitoring threshold of the three-dimensional bridge model based on the analysis error.

[0143] The same load condition is applied to the three-dimensional model of the bridge, and the simulated displacement data of the three-dimensional model of the bridge is obtained by using a displacement change algorithm based on monitoring thresholds, specifically:

[0144] It reads static and moving load information, automatically converts it into a recognizable command stream, and achieves integrated, online modeling and solving of static and moving loads by programmatically editing and rewriting the load mode segments in the input script.

[0145] For static loads, automatically generate or update load patterns and nodal load lists, apply the concentrated force / moment of a node to the structure, and obtain the equilibrium displacement and internal forces through static step calculation. Data items: node number, force component and moment component.

[0146] For moving loads, the system discretizes vehicle axle load, speed, and travel path into time steps or position steps, constructs a load scheduling table that migrates on the node sequence over time or steps, and performs the solution step by step. When the user modifies the load, the system only rewrites the load-related segments and triggers incremental solution, achieving efficient, rollbackable, and observable moving load analysis.

[0147] The moving load includes a basic parameter configuration module, a load editor, a moving load scheduler algorithm, a solver orchestrator, and a result streaming service. The load editor is used for incremental script modification and supports default rollback and versioning. The solver orchestrator has modules for static stepping, error monitoring, failure retry, and rollback, updating the model state only when necessary. The result streaming service pushes displacement and internal force data to the front end in real time.

[0148] The specific implementation method for the moving load is as follows:

[0149] 1) Parameter configuration:

[0150] A. The lane path is discretized into nodes, forming a path node set pathNodes=[n0, n1, ..., n K ]; where n0, n1, ..., n K These are the nodes arranged in order;

[0151] B. Construct the vehicle-axle group parameter set axles=[{axleId,weight_N,axleSpacing,laneOffset,timeOffset}];

[0152] Where: axleId represents the axle number; weight_N represents the axle weight; axleSpacing represents the wheelbase; laneOffset represents the lane offset; and timeOffset represents the time offset.

[0153] C. Global timing parameters: step size dt, vehicle speed v, and node spacing dx;

[0154] D. The loading list for each analysis step k is Schedule[k] = [{nodeId, Fz, Fx, Fy, Mx, My, Mz,laneId, axleId}];

[0155] Where: nodeId is the node number, Fz is the force in the Z direction, Fx is the force in the X direction, Fy is the force in the Y direction, Mx is the torque about the X axis, My is the torque about the Y axis, Mz is the torque about the Z axis, laneId is the lane number, and axleId is the axle number.

[0156] E. Generation principle: Each node corresponds to a valid node length;

[0157] For each lane and each axle, calculate the length of the effective node to which it belongs in step k, and then determine the node to which it belongs. Write the relevant parameters into the schedule[k] for this step; multiple lanes / multiple axles are naturally superimposed in the same step k.

[0158] 2) Implementation of the moving load scheduler algorithm:

[0159] A. Dynamic position calculation algorithm for multi-axle vehicle loads:

[0160] a. During vehicle movement, based on the vehicle speed v and the cumulative wheelbase offset offset_j of each axle j relative to the front axle, the position of each axle j at step k is obtained. ;

[0161] b. Determine the node to which each axle j belongs based on its position x_axle(k) at step k;

[0162] c. Within the range of axle j from the initial step k_start of the effective node length entering node to the final step k_end of the effective node length leaving node, the total cumulative load F_total of axle j acting on node is obtained using the continuous-time integral formula for moving load:

[0163] F_total=

[0164] in: Let be the load value of axle j at step k, which is a constant;

[0165] Let be a position function, representing the load value of axle j at step k in spatial position. The centralizing effect of the location;

[0166] exist When the value is infinity, and at other locations it is 0. The integral represents the load acting at the location. ;

[0167] b. Dynamic response of the structure under moving loads:

[0168] u(k) =

[0169] The meaning is: the dynamic response u(k) of each step k, that is, the displacement of the structure at step k due to the moving load, is the sum of the contributions of all historical loads in the current step k within the time range from 0 to step k.

[0170] Each step within the time range of 0 to step k ,step load The contribution of step k is determined by the load. The size of the distance and the time interval of step k are determined by the magnitude of the distance.

[0171] Let be the unit impulse response function, representing the response at step . The response generated at step k when a unit pulse load is applied.

[0172] This invention also provides an online finite element bridge monitoring and modeling system, comprising:

[0173] The conversion module is used to convert the bridge MidasMCT finite metadata file into an OpenSeesTCL finite metadata file;

[0174] The parsing module is used to enable the browser to load the OpenSeesTCL finite metadata file and parse the OpenSeesTCL finite metadata file into a visualized 3D model of the bridge.

[0175] The load loading and displacement monitoring module is used to access the load conditions applied to the bridge and the monitoring displacement data caused by the bridge under the applied load conditions in real time through the bridge monitoring data interface.

[0176] The bridge 3D model simulation module is used to apply the same load condition to the bridge 3D model, obtain the simulated displacement data of the bridge 3D model through a displacement change algorithm based on a monitoring threshold, compare the monitored displacement data and the simulated displacement data to obtain the analysis error, and adjust the monitoring threshold of the bridge 3D model based on the analysis error.

[0177] The following is an example:

[0178] This embodiment provides an online finite element bridge monitoring finite element modeling system, combined with... Figure 2 ,include:

[0179] Parse the limited metadata file of the bridge MidasMCT:

[0180] Using MidasMCT finite metadata files as input, multimodal analysis and semantic understanding are performed on component types, section symbols, dimension annotations, material identifiers, connection relationships, load symbols, and paths. Through topology reconstruction, component axes are transformed into node-element networks, forming a unified data platform (containing a structured dictionary of units, materials, sections, nodes, elements, constraints, and loads). Based on this, materials and sections undergo normalized mapping and automatic modeling: on the material side, material properties are distinguished and unit-consistent conversion and elastic modulus mapping are performed; on the section side, database sections, variable sections, and numerical sections (including outer contours and multi-hole point sets) are identified and classified, automatically generating fiber sections and end-consistent modeling; on the connection side, rigid and elastic connections are analyzed, and materials and coefficients are generated according to degrees of freedom; on the load side, static and moving loads are discretized into node-step scheduling tables. Finally, this is automatically converted into a command stream recognizable by OpenSees. Currently, it supports modeling of beam bridges, cable-stayed bridges, and suspension bridges from MidasMCT finite metadata files to OpenSees. The specific analysis approach consists of the following parts:

[0181] 1. Unified data platform and semantic parsing:

[0182] 1) Unified data structure dictionary: data = { 'units', 'materials', 'sections', 'nodes', 'elements', 'constraints', 'loads', 'elastic_link', 'rigid_link', 'line_lanes', 'nodal_mass', 'analysis'}. The parsing results from each module are incorporated into this unified dictionary, forming a "structured middle platform".

[0183] 2) Multimodal analysis:

[0184] A. Unit: Analytical * UNIT (force, length);

[0185] B. Materials: Parse *MATERIAL (USER / CONC / STEEL) and perform semantic mapping on common concrete / steel materials;

[0186] C. Section: Analytical *SECTION (DBUSER / TAPERED / PSCVALUE / PSCPSC), including outer contour and multiple internal hole point sets;

[0187] D. Nodes and Elements: Analyze *NODE and *ELEMENT to construct a node-element network;

[0188] E. Connection: Resolving *ELASTICLINK and *RIGIDLINK;

[0189] F. Constraints: Parse *CONSTRAINT (supports range and step expressions, such as 1to10by2);

[0190] G. Load and Lane: Analyze *CONLOAD and *LINELANE(CH) to form a static / dynamic load scheduling table.

[0191] 2. Unit unification and dimensional mapping:

[0192] Set the force unit conversion factor to kF and the length to kL. The program will automatically map from MCT units to OpenSees requirements (by default, N and mm are the target).

[0193] The program has a built-in `convert_units(value, from_unit)` function to perform a unified conversion on the above quantities, ensuring that engineering data from different sources can be directly applied to OpenSees.

[0194] 3. Topology Reconstruction and Local Coordinates

[0195] From axis to network: Read *NODE and *ELEMENT, establish node coordinate table and cell connectivity, and form a unified topology.

[0196] Geometric transformation generation using geomTransf:

[0197] 1) Local x-axis (unit vector): x_raw = (xj - xi, yj - yi, zj - zi); x = x_raw / ||x_raw||

[0198] 2) The reference vector is chosen (not collinear with x_raw) to obtain the vector v_raw. In calculating the local z-axis (unit vector): z_raw = x × v; z = z_raw / ||z_raw||

[0199] 3) Output OpenSees command: geomTransf Linear tag zx_x zx_y zx_z (i.e., the z_raw component above)

[0200] For large angles or special directions, the program ensures numerical stability and non-collinearity by using candidate reference vectors and cross products.

[0201] 4. Constraint Resolution

[0202] 1) Supports "range + step" constraints (e.g., 1to10by2) and mask formats (e.g., 001000).

[0203] 2) Stability check and auxiliary constraints: When constraints are significantly insufficient, automatic prompts are given and complete constraints can be added at the lowest Z coordinate node;

[0204] 3) Zero-length cell filtering: If the coordinates of the two ends are close to the same position (threshold such as 1e-6), it will be automatically skipped and a prompt will be given.

[0205] Modeling of different bridge cross sections:

[0206] It can automatically convert cross-section information read from MidasMCT finite element metadata files into a command stream recognizable by OpenSees, and supports multiple cross-section types, suitable for parametric and fiberized modeling on finite element platforms. The specific process involves parsing the cross-section definition from the engineering data file, extracting the cross-section type and shape parameters, and performing regular mapping and classification processing. Separate modeling processes are established for user-defined and variable cross-section types (such as database / user DBUSER, variable cross-section TAPERED, numerical cross-section PSCVALUE, design variable cross-section PSCTAPERED, and design numerical cross-section PSCPSC). The specific generation process consists of the following parts:

[0207] 1. DBUSER / TAPERED: Identifies the type and symbol, mapping to fiber section or numerical section.

[0208] Rectangular fiber patch: patch rect mat nx ny x1 y1 x2 y2. The program automatically determines nx and ny based on the width, height, and desired mesh size.

[0209] Circular / ring-shaped fibers: Outer diameter and thickness are converted into inner and outer radii to form ring-shaped fibers (or semi-rings, divided by angle segments).

[0210] I-beam / track / box type: Composed of several rectangular / circular arc segments, automatically positioned and assembled in parallel to form a complete cross section.

[0211] 2. Numerical Section (PSCVALUE / PSCPSC):

[0212] The complex cross-section of "polygon + multiple cavities" is stably mapped to OpenSees Fiber, strictly maintaining area consistency.

[0213] 3. Consistent modeling of variable cross sections:

[0214] Equivalent interpolation along beam length for variable cross-section (integral at both ends - fixed position)

[0215] The local coordinates of the principal axes of the beam element are ξ ∈ [0, 1].

[0216] Linear equivalent interpolation is performed on the cross-sectional properties at both ends, P0 (starting end) and P1 (ending end): P(ξ) = (1 − ξ)· P0 +ξ·P1

[0217] The beam uses forceBeamColumn + beamIntegration FixedLocation for interpolation at both ends of the cross-section: beamIntegration FixedLocation intTag 2 secStart secEnd 0 1; elementforceBeamColumn ele ij transf intTag.

[0218] For various geometries, including solid-web rectangles (SB), circular / tubular (P / SR), I-beams (H), solid-web track-type (STRK), semi-solid-web track-type (HTRK), and internal octagonal box-type (ROCT), a material-section association lookup module binds material numbers to section parameters, calls the corresponding section generation module to construct fiber sections, and automatically outputs the section definition as a structural analysis script. For numerical and variable sections of prestressed concrete, fiber meshing is constructed based on outer contour points and multi-internal hole contour sets, supporting consistent modeling representation of the start and end points; it also provides anomaly alarms and degradation output strategies to ensure that compatible section definitions can still be generated even when parameters are incomplete or parsing fails. The device consists of a data parsing and parameter extraction module, a material mapping module, a shape recognition and strategy distribution module, a fiber mesh generation module, a script writing module, and an anomaly rollback module, forming an scalable shape plug-in mechanism and a robust engineering implementation process. Its features are summarized as follows:

[0219] Automatically identifies various bridge cross-sections and generates fiber cross-sections with a single click, reducing errors and time costs associated with manual modeling. It provides a unified parametric representation of cross-section types, ensuring consistent modeling at both ends of variable cross-sections. It supports complex box-type, track-type, and multi-hole cross-sections, refining the accuracy of stress simulation. It features error alerts and degradation rollback, improving the robustness and maintainability of the engineering process. Its modular design facilitates the addition of new cross-section shapes and material models, offering excellent scalability.

[0220] Modeling of different bridge material properties:

[0221] It can automatically convert bridge material information read from MidasMCT finite metadata files into a command stream recognizable by OpenSees, and normalize and map the material definitions in the engineering data into material constitutive models that can be executed by the analysis solver. The specific process is as follows: read the material number and attribute set from the dataset, classify and process them according to material type and perform unit consistency conversion, and distinguish between rigid connections and general elastic connections for connection materials.

[0222] Real-time dynamic modification of bridge static and moving loads for analysis:

[0223] It can read static and moving load information from MidasMCT finite metadata files and automatically convert it into a command stream recognizable by OpenSees. By programmatically editing and rewriting the load mode segment in the OpenSees input script, it can achieve integrated, online modeling and solving of static and moving loads.

[0224] 1. Automatically generate or update load patterns and nodal load lists for static loads. Apply the concentrated force / moment of a node to the structure and obtain the equilibrium displacement and internal forces through static step calculation. Data items: node number nodeId; force components Fx, Fy, Fz; moment components Mx, My, Mz (units recommended N, N·mm), such as: pattern Plain 1 Linear{load 106 0.0 0.0 -8.0e4 0.0 0.0 0.0};

[0225] 2. For moving loads, the system discretizes them into time steps (or position steps) based on vehicle axle load, speed, and travel path. It constructs a "load scheduling table" that migrates across the node sequence over time (or steps) and executes the solution step by step. When a user modifies the load, the system only rewrites the "load-related segments" and triggers "incremental solution," achieving efficient, rollback-capable, and observable moving load analysis. Moving load analysis consists of four main modules: a load editor (incrementally modifies the OpenSees script's timeSeries / pattern / load segment, supporting default rollback and versioning), a moving load scheduler algorithm, a solver orchestrator (static stepping, error monitoring, failure retry and rollback; updating the model state only when necessary), and a result streaming service (pushing displacement and internal force data to the front end in real time). The specific implementation steps of these four modules are as follows:

[0226] 1,

[0227] 1) Implementation of vehicle axle load, speed, and time step.

[0228] A. Lane path (discrete): pathNodes = [n0, n1, ..., nK] (unit: mm)

[0229] B. Vehicle-Axle Group: axles = [{id, weight_N, axleSpacing_mm, laneOffset_mm, timeOffset_s}, ...]

[0230] C. Global timing parameters: dt (step size s), v (vehicle speed mm / s), node spacing dx (mm)

[0231] D. The formula for the number of steps per node is s = ⌊dx / v⋅dt⌋

[0232] E. The activation start step formula for the i-th node is ti = i ⋅ s

[0233] F. The formula for the time offset step of the a-th root axis is Δa = ⌊timeOffseta / dt⌋

[0234] G. The formula for converting axis spacing to node offset (if using "jump with node") is δia≈⌊axleSpacinga / dx⌋

[0235] H. Loading list for each analysis step k: Schedule[k] = [{nodeId, Fz, Fx, Fy, Mx, My,Mz, laneId, axleId}, ...]

[0236] I. Generation principle: For each lane and each axle, calculate its corresponding node position in step k and write it into the list for that step; multiple lanes / multiple axles are naturally superimposed within the same step.

[0237] 2) Implementation of the moving load scheduler algorithm

[0238] A. Dynamic position calculation algorithm for multi-axle vehicle load: a. Time evolution formula for axle position: x_axle(t) = v × dt - offset_i, where x_axle(t) is the position of the i-th axle at time t, v is the vehicle speed (mm / s), and offset_i is the cumulative offset of the i-th axle relative to the front axle. b. Calculation of cumulative axle offset: offset_i = Σ_{j=0}^{i-1} d_j (i > 0), offset_0 = 0, where d_j is the j-th axle offset, and offset_i is the cumulative offset of the i-th axle.

[0239] B. Vehicle load node allocation algorithm: a. Nearest node search algorithm: node_target = argmin_{k}|x_k - x_axle(t)|, where x_k is the x-coordinate of the k-th node, x_axle(t) is the current position of the axle, and node_target is the nearest node number. b. Load application boundary condition determination: load_condition = (x_axle(t) ≥ 0) ∧ (x_axle(t) ≤ L_total), where L_total is the total length of the bridge. The load is applied when the condition is met; otherwise, it is not applied.

[0240] C. Continuous-time integral formula for moving load: a. Integral expression of load application time: F_total(t) = ∫_{t_start}^{t_end} Σ_{i=1}^{n_axles} F_i(τ) × δ(x_i(τ)) dτ, where F_i(τ) is the load of the i-th axle at time τ, δ(x_i(τ)) is the position function (1 when the axle is on the bridge, 0 otherwise), and t_start, t_end are the load application time ranges. b. Time history of dynamic response: u(t) = ∫_{0}^{t} G(t-τ)× F_total(τ) dτ, where u(t) is the structural response, G(t-τ) is the structural transfer function, and F_total(τ) is the total load time history.

[0241] 3) Implementation of the load editor (Incremental update of OpenSees fragments)

[0242] A. Segment structure (once per step, statically progressive): timeSeries Constant tsId, pattern PlainpatId tsId { load node ...} (loads the corresponding nodes of all lanes and axes within the same step), analyze 1, remove loadPattern patId, remove timeSeries tsId

[0243] B. Incremental Strategy: Do not modify the model, elements, or constraint segments; only rewrite the fragments of "time series + load mode + loading step" and maintain "load segment hash / version number"; after user changes, only replace the fragments of the affected steps (supports step range recalculation); when the schedule table is empty or abnormal, write an empty step (do not load) or roll back to the last valid version.

[0244] 4) Solve the programmable structure (robust static step and rollback)

[0245] A. Typical static configuration (illustration):

[0246] constraints Transformation;

[0247] Numberer AMD (or RCM);

[0248] system UmfPack (or BandGeneral);

[0249] testNormDispIncr tol iters (such as 1e-6, 50);

[0250] algorithm Newton / KrylovNewton / Linear;

[0251] Integrator LoadControl 1.0 (steps through the loads of the current step, each time "full").

[0252] analysis Static;

[0253] B. In-step process (step k): In-step process (step k): Write: timeSeries Constant + patternPlain {load...} (containing step k positions for all lanes / axles); Execute: analyze 1 (check return codes and residuals); Clean up code remove loadPattern + remove timeSeries.

[0254] C. Failure Handling and Rollback: Increase the iteration limit, tighten the tolerance, switch algorithms (Newton → KrylovNewton / ModifiedNewton), split a "full-scale loading within a step" into "sub-steps" (e.g., execute LoadControl 0.5 twice), and then clean up. If it still fails, undo the current step, record the error context, maintain the "converged state" of the previous step, and continue or interrupt.

[0255] 5) Result streaming service (real-time observable)

[0256] A. Recorder configuration (illustrated): recorder Node -file results / node_disp.out -time -node ... -dof ... disp (displacement); recorder Element -file results / ele_force.out -time -ele ... force (internal force).

[0257] B. Streaming Pipeline: The backend process appends to the tail listener file; parses it into structured events (step number, time, entity, value); pushes them to the frontend via WebSocket; supports reconnection after disconnection, rate limiting, and merging (in batches of 50–100 ms).

[0258] When a load change is detected, the device only rewrites the load-related segments and triggers incremental solving, avoiding repeated construction of the overall model and ensuring a low-latency closed loop of "editing-solving-visualization". Its characteristics are summarized as follows:

[0259] Real-time performance: Load editing and result feedback are linked in a streaming manner, supporting second-level feedback and interactive verification of static and moving loads.

[0260] Accuracy: Moving loads are migrated to discrete nodes in time / location steps, maintaining the vehicle wheelbase, speed and force transmission consistent with the bridge deck mesh.

[0261] High efficiency: Only incremental modifications are made to the load segment and the solution is obtained step by step, which significantly reduces computation and I / O overhead.

[0262] Ease of use and scalability: A unified load description and scheduling interface facilitates expansion to multi-vehicle platooning, lane allocation, time-varying coefficients, and combined load case analysis.

[0263] Convert the OpenSees TCL finite metadata file into a 3D model of the bridge in the browser:

[0264] This system enables real-time parsing, modeling, and 3D visualization of OpenSees TCL limited metadata files for browser-based applications. It consists of a front-end parser and a lightweight server. Without altering the original TCL semantics, it parses model text online and constructs 3D scenes. The parser performs line-by-line semantic recognition and mapping of model statements, including nodes, elements, materials, and sections. At the section level, it supports block-level parsing and geometric construction of fiber sections. Geometry generation utilizes Three.js: based on the section shape, it extrudes solids according to element lengths. The system achieves an online closed loop on the browser side: "loading TCL text—parsing—building geometry—assembling elements," enabling rapid reproduction and verification of engineering-grade bridge models without the need for a local analyzer. Its features are summarized below:

[0265] (1) Ready to use: The bridge model is visualized online by loading the limited metadata files of OpenSees TCL in the pure front end, without the need for local plugins and complex deployment.

[0266] (2) High-fidelity geometry: native support fiber cross-section (circle / rectangle / polygon and inner hole) and polymer cross-section assembly, the cross-section shape and hole are completely restored.

[0267] (3) Strong compatibility: Adapts to commonly used OpenSees modeling statements and annotation / record filtering logic, with robust fault tolerance.

[0268] (4) Extensible: Patch and polygon are used as plug-in shape entry points, which makes it easy to add new cross sections and attribute rules, and is suitable for online reproduction and teaching demonstration of various types of bridge components.

[0269] This embodiment describes an online finite element bridge monitoring and modeling system and method. The system can import and process MidasMCT finite element metadata files and perform finite element metadata analysis using converted OpenSees TCL finite element metadata files. The main advantages of this embodiment include enhanced visualization, improved processing efficiency, and user-friendliness.

[0270] Constituent elements:

[0271] Hardware components: This system includes a computer with a high-performance processor and large memory to process bridge finite element data. In addition, a data interface for bridge monitoring and on-site displacement monitoring sensors are required.

[0272] Software Application: The system runs a specific online bridge finite element analysis application. This application has a user-friendly interface that allows users to import bridge MidasMCT finite element metadata files.

[0273] Data Processing Module: The data processing module is responsible for converting the imported bridge MidasMCT limited metadata file into an OpenSees TCL limited metadata file.

[0274] Operating steps:

[0275] 1. The user opens the bridge real-time static and moving load analysis system.

[0276] 2. Select the bridge MidasMCT limited metadata file and import it into the system.

[0277] 3. The background process converts the bridge's MidasMCT limited metadata file into an OpenSees TCl limited metadata file.

[0278] 4. The browser parses the OpenSees TCl limited metadata file into a visualized 3D bridge model.

[0279] The present invention has the following advantages:

[0280] 1. This invention proposes a finite element transformation method based on OpenSees, which can be implemented via the Internet.

[0281] The line processing converts MidasMCT finite metadata files to OpenSees TCL finite metadata files, enabling the conversion of bridge structure nodes, sections, and lines.

[0282] 2. Through the analysis of static and moving loads, the displacement calculation of key bridge components under different working conditions was realized using OpenSees TCL finite element metadata files. Compared with the commercial finite element software Midas, due to the different algorithms, the error of simply supported beam bridges was controlled within 5%, and the error of large and complex bridges was controlled within 10%.

[0283] 3. By using lightweight web technology, the limited metadata files of OpenSees TCL were loaded in the browser, enabling lightweight online access.

[0284] 4. By accessing the real-time monitoring displacement data in bridge monitoring, it is possible to compare it with the simulation data of the finite element method of this invention, further verifying the accuracy of the finite element simulation model, so as to adjust the bridge monitoring threshold.

[0285] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. An online finite element bridge monitoring and modeling method, characterized in that, include: Step S1: Convert the bridge MidasMCT finite metadata file into an OpenSeesTCL finite metadata file; Step S2: The browser loads the OpenSeesTCL finite metadata file and parses the OpenSeesTCL finite metadata file into a visualized 3D model of the bridge. Step S3: Real-time access to the load conditions applied to the bridge and the monitoring displacement data caused by the applied load conditions through the bridge monitoring data interface. Step S4: Apply the same load condition to the three-dimensional model of the bridge, and obtain the simulated displacement data of the three-dimensional model of the bridge by using a displacement change algorithm based on monitoring thresholds; The monitored displacement data and the simulated displacement data are compared to obtain the analysis error; based on the analysis error, the monitoring threshold of the bridge three-dimensional model is adjusted.

2. The online finite element bridge monitoring and modeling method according to claim 1, characterized in that, Step S1 is as follows: Step S11: Establish a unified data structure; Step S12: Perform multimodal parsing and semantic understanding on the bridge's MidasMCT finite metadata file. Through topology reconstruction, convert each bridge component along its axis into node-element units. Further analysis yields the following attributes for each element: element size unit, element force unit, element material, element cross-section, node number within the element, constraints of the element, loads applied to the element, and connection method between elements. Store all attributes of each parsed element into the unified data structure. After all elements have been parsed, a structured dictionary is obtained, forming a structured middleware platform. Step S13: Perform unit consistency and dimension mapping transformation on the data stored in the structured dictionary to obtain the transformed structured dictionary; Step S14: Based on the transformed structured dictionary, perform topology reconstruction transformation and local coordinate reconstruction of each element, perform consistent modeling and expression of the cross section and modeling and expression of material properties, and obtain the OpenSeesTCL finite metadata file.

3. The online finite element bridge monitoring and modeling method according to claim 2, characterized in that, Multimodal parsing of the bridge's MidasMCT finite metadata file includes: Standardized mapping and automatic modeling of materials and cross sections: On the material side, material properties are distinguished and unit-consistent conversion and elastic modulus mapping are performed; the material properties include concrete and steel materials; on the cross section side, database cross sections, variable cross sections and numerical cross sections are identified and classified, and the cross sections include outer contours and multiple internal hole point sets; the identified cross sections are automatically generated to generate fiber cross sections and end-consistent modeling. The connection side analyzes rigid and elastic connection methods, and generates material parameters and coefficients according to the degree of freedom of the connection method. The static load and moving load are discretized into a node-step scheduling table on the load side.

4. The online finite element bridge monitoring and modeling method according to claim 2, characterized in that, The data stored in the structured dictionary is subjected to unit consistency and dimensional mapping transformation, specifically as follows: Force unit conversion factor Conversion factor for length units The following formula is used for unit unification and dimensional mapping transformation: ； in: , , , , , These are the converted force, length, area, moment of inertia, elastic modulus, and elastic coefficient, respectively. , , , , , These represent the force, length, area, moment of inertia, elastic modulus, and elastic coefficient before conversion, respectively; the unit for force after conversion is Newtons (N); and the unit for length after conversion is millimeters (mm).

5. The online finite element bridge monitoring and modeling method according to claim 2, characterized in that, The topology reconstruction transformation and the local coordinate reconstruction of each cell are specifically as follows: Establish a global coordinate system; transform each type of bridge component into a node-element along its axis, and create a three-dimensional coordinate table of nodes in the global coordinate system; establish the connectivity between nodes and elements to form a unified topology; For each element in the global coordinate system, a local coordinate system for each element is established. The method for establishing the local coordinate system is as follows: The three-dimensional global coordinates of the two nodes connected by the unit are: First node Second node Then define the local x-axis of the local coordinate system as x_raw = x = x_raw / ||x_raw||; Specify a reference vector v_raw that is not collinear with the local x-axis x_raw, and calculate the local z-axis z_raw = x_raw × v_raw; z = z_raw / ||z_raw||. This leads to the establishment of a local coordinate system for the element.

6. The online finite element bridge monitoring and modeling method according to claim 2, characterized in that, The consistent modeling and representation of the cross-section specifically includes: Analyze the cross-sectional parameters to extract the cross-sectional type and shape parameters; Based on the cross-section type and cross-section shape parameters, regular mapping and classification processing are performed: ①The cross-section type is a numerical cross-section: If the cross-section type is a numerical cross-section, calculate the outer contour area A_outer and the area of ​​each inner hole; subtract the outer contour area A_outer from the area of ​​all inner holes to obtain the effective cross-section area A_eff; Use the ray casting method to determine the number of valid sampling points N_valid on the outer contour of the cross section that are not in any inner hole; The equivalent area of ​​a single fiber is obtained using the formula A_fiber = A_eff / N_valid; For each valid sampling point, generate fibers of equal area, with the material number inherited from the parent material; This allows the numerical cross-section with multiple cavities to be stably mapped to a fiber cross-section, maintaining area consistency. ②The cross-section type is variable cross-section: The method for consistent modeling of variable cross sections is as follows: The local coordinates of the principal axes of the beam element are ξ∈[0, 1]. For the starting section P0 and the ending section P1 at both ends of the beam, linear equivalent interpolation is performed on the points of the beam element along the beam length to construct a variable cross section with consistent modeling expression at the starting and ending ends. Therefore, for cross-sections with different geometries, the material-section association lookup module binds the material number to the cross-section parameters, calls the corresponding cross-section generation module to construct the fiber cross-section, and automatically outputs the cross-section definition as the structural analysis script. For numerical cross-sections and variable cross-sections of prestressed concrete, fiber subdivision is constructed based on the outer contour points and the set of multiple inner hole contours, supporting consistent modeling expression of the start and end ends. At the same time, anomaly alarms and degradation output strategies are provided to ensure that compatible cross-section definitions can still be generated when parameters are incomplete or parsing fails. Specifically, it includes a data parsing and parameter extraction module, a material mapping module, a shape recognition and strategy distribution module, a fiber mesh generation module, a script writing module, and an exception rollback module, forming an scalable shape plugin mechanism and a robust engineering implementation process.

7. The online finite element bridge monitoring and modeling method according to claim 2, characterized in that, The modeling and expression of material properties are specifically as follows: From the transformed structured dictionary, the material definition is read, and the material definition is normalized and mapped to the material constitutive model that can be executed by the analysis solver. The specific process is as follows: read the material number and the set of material properties, classify and process them according to the material type and convert them to unit consistency, and distinguish between rigid connection and general elastic connection for connection materials. Output the elastic material definition containing Poisson's ratio for user-defined materials, read the elastic modulus and Poisson's ratio, and unify them to N / mm²; For concrete and steel, the standard elastic modulus is used, and the grade is mapped to the target elastic modulus. If the connecting materials are general elastic connections, determine the start node and end node in the unit formed by the general elastic connection, the material labels corresponding to each of the 6 degrees of freedom, generate an independent uniaxial material in each degree of freedom direction, and use the material with E=0 if the stiffness is 0. If the connecting materials are rigid, the large E approximation is used. In the element formed by the connecting materials, the master node and slave node are determined. Through multi-point constraint commands, the slave node and the master node are forced to maintain the same motion in the six degrees of freedom that need to be constrained, thus obtaining the modeled element of the connecting materials. The six degrees of freedom include X translation, Y translation, Z translation, rotation about X, rotation about Y, and rotation about Z.

8. The online finite element bridge monitoring and modeling method according to claim 1, characterized in that, In step S4, the same load condition is applied to the three-dimensional model of the bridge, and the simulated displacement data of the three-dimensional model of the bridge is obtained through a displacement change algorithm based on a monitoring threshold. Specifically: It reads static and moving load information, automatically converts it into a recognizable command stream, and achieves integrated, online modeling and solving of static and moving loads by programmatically editing and rewriting the load mode segments in the input script. For static loads, automatically generate or update load patterns and nodal load lists, apply the concentrated force / moment of a node to the structure, and obtain the equilibrium displacement and internal forces through static step calculation. Data items: node number, force component and moment component. For moving loads, the system discretizes vehicle axle load, speed, and travel path into time steps or position steps, constructs a load scheduling table that migrates on the node sequence over time or steps, and performs the solution step by step. When the user modifies the load, the system only rewrites the load-related segments and triggers incremental solution, achieving efficient, rollbackable, and observable moving load analysis. The moving load includes a basic parameter configuration module, a load editor, a moving load scheduler algorithm, a solver orchestrator, and a result streaming service; wherein, the load editor is used to incrementally modify the script and supports default rollback and versioning; The solver orchestrator has modules for static stepping, error monitoring, failure retry and rollback, and updates the model state only when necessary; the result streaming service is used to push displacement and internal force to the front end in real time.

9. The online finite element bridge monitoring and modeling method according to claim 8, characterized in that, The specific implementation method for the moving load is as follows: 1) Parameter configuration: A. The lane path is discretized into nodes, forming a path node set pathNodes=[n0, n1, ..., n K ]; where n0, n1, ..., n K These are the nodes arranged in order; B. Construct the vehicle-axle group parameter set axles=[{axleId,weight_N,axleSpacing,laneOffset,timeOffset}]; Where: axleId represents the axle number; weight_N represents the axle weight; axleSpacing represents the wheelbase; laneOffset represents the lane offset; and timeOffset represents the time offset. C. Global timing parameters: step size dt, vehicle speed v, and node spacing dx; D. The loading list for each analysis step k is Schedule[k] = [{nodeId, Fz, Fx, Fy, Mx, My, Mz,laneId, axleId}]; Where: nodeId is the node number, Fz is the force in the Z direction, Fx is the force in the X direction, Fy is the force in the Y direction, Mx is the torque about the X axis, My is the torque about the Y axis, Mz is the torque about the Z axis, laneId is the lane number, and axleId is the axle number. E. Generation principle: Each node corresponds to a valid node length; For each lane and each axle, calculate the length of the effective node to which it belongs in step k, and then determine the node to which it belongs. Write the relevant parameters into the schedule[k] for this step; multiple lanes / multiple axles are naturally superimposed in the same step k. 2) Implementation of the moving load scheduler algorithm: A. Dynamic position calculation algorithm for multi-axle vehicle loads: a. During vehicle movement, based on the vehicle speed v and the cumulative wheelbase offset offset_j of each axle j relative to the front axle, the position of each axle j at step k is obtained. ; b. Determine the node to which each axle j belongs based on its position x_axle(k) at step k; c. Within the range of axle j from the initial step k_start of the effective node length entering node to the final step k_end of the effective node length leaving node, the total cumulative load F_total of axle j acting on node is obtained using the continuous-time integral formula for moving load: F_total= ; in: Let be the load value of axle j at step k, which is a constant; Let be a position function, representing the load value of axle j at step k in spatial position. The centralizing effect of the location; exist When the value is infinity, and at other locations it is 0. The integral represents the load acting at the location. ; b. Dynamic response of the structure under moving loads: u(k) = ; The meaning is: the dynamic response u(k) of each step k, that is, the displacement of the structure at step k due to the moving load, is the sum of the contributions of all historical loads in the current step k within the time range from 0 to step k. Each step within the time range of 0 to step k ,step load The contribution of step k is determined by the load. The size of the distance and the time interval of step k are determined by the magnitude of the distance. Let be the unit impulse response function, representing the response at step . The response generated at step k when a unit pulse load is applied.

10. A system for implementing the online finite element bridge monitoring and modeling method according to any one of claims 1-9, characterized in that, include: The conversion module is used to convert the bridge MidasMCT finite metadata file into an OpenSeesTCL finite metadata file; The parsing module is used to enable the browser to load the OpenSeesTCL finite metadata file and parse the OpenSeesTCL finite metadata file into a visualized 3D model of the bridge. The load loading and displacement monitoring module is used to access the load conditions applied to the bridge and the monitoring displacement data caused by the bridge under the applied load conditions in real time through the bridge monitoring data interface. The bridge 3D model simulation module is used to apply the same load conditions to the bridge 3D model and obtain the simulated displacement data of the bridge 3D model through a displacement change algorithm based on monitoring thresholds. The monitored displacement data and the simulated displacement data are compared to obtain the analysis error; based on the analysis error, the monitoring threshold of the bridge three-dimensional model is adjusted.

Images