An agent-based cable force tensioning decision method and system
Patent Information
- Application Number
- CN202611273822.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-21
- Publication Date
- 2026-10-30
AI Technical Summary
第一，有限元模型通常基于设计参数建立，难以及时反映现场支承刚度、节段重量、临时构造、温度场、锚具回缩和设备误差等实际状态；
将有限元模型实体与现场测点、张拉设备和施工阶段建立显式关联，使现场数据能够按映射关系投影到有限元模型对应节点、单元和自由度；通过智能体内的数据处理功能进行质量检查、时空对齐和温度修正，减少因数据源异构、采样不同步或测点异常导致的错误张拉决策，通过智能体内的模型修正功能依据现场数据修正施工阶段有限元模型，使张拉方案基于当前真实结构状态而非静态设计模型，通过智能体内的方案生成功能利用索力影响矩阵或有限元响应面生成候选张拉方案，降低人工反复试算成本，通过智能体内的方案校核功能逐级复算候选张拉方案，能够同时约束索力、线形、索塔偏位、应力、相邻索影响和设备能力，通过智能体内的执行反馈功能将张拉方案转换为现场指令并接收复测反馈，在每级张拉后修正后续方案，降低调索风险并提高成桥状态控制精度。
Smart Images
Figure CN123013561A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of bridge construction control technology, and in particular relates to a cable tensioning decision-making method and system based on intelligent agents. Background Technology
[0002] Cable-stayed bridges, suspension bridges, tied arch bridges, spatial cable net structures, and cable membrane structures require tensioning or cable adjustment during construction to gradually bring the structure closer to the design target state. Cable tensioning not only affects the stress of individual cables but also changes the overall bridge alignment, internal forces, and stability state through the main beam, towers, arch ribs, anchorage zones, and temporary supports. For single-tower cable-stayed bridges, spatial torsion cable-stayed bridges, irregular steel tower bridges, and in cases of support removal or system conversion, the coupling effect between cables is more pronounced. Tensioning deviations of individual cables may lead to changes in the cable forces of adjacent cables, increased tower displacement, or deviations in the main beam alignment from the design target. Current tension control methods typically rely on finite element analysis to pre-calculate the target cable force, followed by manual adjustments on-site based on cable force, alignment, or experience. However, this approach has the following drawbacks: First, finite element models are usually built based on design parameters, which makes it difficult to reflect the actual conditions on site, such as support stiffness, segment weight, temporary structures, temperature field, anchor retraction, and equipment errors. Second, the field data comes from multiple sources such as cable tensioners, tensioning equipment, total stations, levels, strain gauges, temperature sensors, and construction records. The physical quantities, coordinate systems, sampling frequencies, and construction stages of the data are inconsistent, making it difficult to directly form a calculable model verification relationship with the finite element model. Third, manual calculation and cable adjustment mainly target the current cable or local indicators, making it difficult to simultaneously consider subsequent construction stages, the impact of adjacent cables, equipment capacity, construction sequence, and safety thresholds. Fourth, traditional software often provides finite element analysis or data display functions, but cannot convert field data into data that can be used for tension calculations based on the model entity relationships. Summary of the Invention
[0003] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a cable tensioning decision-making method and system based on intelligent agents. It uses a finite element construction stage model as the basis for mechanical calculation, multi-source data from the site as state feedback, and internal functional arrangement of intelligent agents as the decision-making mechanism. This enables automated processing of monitoring data, real-time correction of the finite element model, intelligent generation of tensioning schemes, hierarchical verification, instruction issuance and closed-loop iterative updates, thereby improving the accuracy and intelligence level of tensioning construction control for cable-stayed bridges.
[0004] The objective of this invention is achieved through the following technical solution: A cable tensioning decision-making method and system based on intelligent agents includes the following steps: Step S1: Input basic data into the database, obtain the finite element model of the target cable-bearing structure, the target bridge design, the construction stage division and tensioning equipment parameters, extract cable elements, main beam nodes, tower nodes, anchorage zone elements, temporary supports, construction loads and boundary conditions from the finite element model, and form a model entity table containing model entities, spatial location, component type, construction stage and mechanical degrees of freedom; Step S2: Obtain the field data channel table and establish the relationship between the field data and the model entity table; Step S3: Intelligent agent data processing. The intelligent agent reads the field data and performs channel integrity checks, abnormal transition checks, timestamp alignment, temperature compensation, equipment calibration coefficient conversion, and construction stage matching on the field data to obtain the field state vector of the current construction stage. Step S4: Intelligent agent model correction. The intelligent agent projects the field state vector to the corresponding node, element or degree of freedom of the finite element model according to the relationship between the field data and the model entity table, calculates the residual between the field state vector and the finite element predicted state, and corrects the model parameters of the finite element model based on the residual to obtain the corrected finite element model for the current construction stage. Step S5: The agent generates candidate tensioning schemes. The agent calls the modified finite element model to apply unit perturbation calculations to the candidate tensioning cables, candidate tensioning sequence and candidate tensioning increment, forming a linear cable force influence matrix or finite element response surface. Then, it calls the tensioning decision knowledge base and uses the completed bridge cable force, main beam alignment, tower deflection, component stress, construction safety and equipment capacity as constraints to generate multiple sets of candidate tensioning schemes through multi-objective optimization. Step S6: Step-by-step finite element scheme verification. The agent performs finite element recalculation on each candidate tensioning scheme to determine whether the cable force deviation, main beam alignment deviation, tower displacement, component stress, anchorage zone stress, adjacent cable influence, jack capacity, and single-stage tensioning increment meet the preset constraints, and outputs the tensioning scheme that passes the verification. Step S7: The intelligent agent outputs on-site execution instructions; Step S8: After tensioning each level or each cable, recollect the field data and return to step S3 to revise the subsequent tensioning scheme.
[0005] In one implementation, in step S2, the association relationship includes a model-site mapping set, and each mapping record in the set includes at least the finite element model node number or element number, cable number, beam end or tower end anchor point number, measuring point channel number, physical quantity type, coordinate transformation parameters, data conversion coefficient, reliability weight, construction stage number, and effective time window.
[0006] In one implementation, in step S3, the field state vector is represented as: ; in, For on-site cable tension, For hydraulic jacks, For the jack stroke, The displacement of the main beam. For Sota's deviation, For component stress, To measure the strain at the point, For the temperature field, For environmental load, This is a status indicator for the construction phase.
[0007] In one implementation, the parameter sensitivity screening method in step S4 is as follows: The parameters of each candidate model are perturbed, and the normalized comprehensive sensitivity coefficient of each parameter to cable force, main beam alignment, tower deviation and component stress is calculated by the central difference method. Low sensitivity parameters with comprehensive sensitivity coefficients lower than the preset threshold are fixed as the initial design values and do not participate in the optimization iteration. Define the residual vector for the construction phase as: ; The finite element model parameters are then corrected using the following objective function: ; in, Let be the residual vector for the t-th construction stage. For the model parameters to be corrected, These are the initial model parameters. For the finite element model of construction phase t, For finite element method or response surface method functions, This is the mapping matrix from the finite element calculation output to the field observations. The observation weight matrix, The regularization coefficient is . The cable force load vector, For construction load vectors, This represents the temperature field of the structure.
[0008] In one embodiment, in step S4, the model parameters to be corrected include at least one of the following: effective elastic modulus of cable, equivalent stiffness correction coefficient for cable sag, anchor retraction, temporary support stiffness, construction load deviation, temperature correction coefficient, weight correction coefficient for main beam segment, boundary condition correction coefficient, and elastic modulus correction coefficient for main beam / tower material.
[0009] In one implementation, in step S5, the linear relationship of the linear cable force influence matrix is as follows: ; The second-order response surface is represented as:
[0010] in, For candidate tension increment vectors, The structural response change vector, Let i be the predicted value of the i-th observed target. Let i be the baseline value for the i-th observed target in the current state. Let be the first-order influence coefficient of the tension increment of the k-th cable on the i-th observed target. Let be the second-order coupling influence coefficient of the tension increments of the k-th and l-th cables on the i-th observed target; The observation targets include at least one of the following: cable force, main beam alignment, tower deviation, component stress, temporary support reaction force, and anchorage zone stress.
[0011] In one implementation, in step S5, the multi-objective optimization function is: ; The constraints include: ; ; ; ,or ; ; ; in, To design the target state; Finite element prediction of candidate schemes; , This is a diagonal weight matrix used for normalizing physical quantities of different dimensions; For tensioning sequence risk indicators; For the number of adjacent cables that need to be adjusted synchronously, It is an index of strain energy or bending moment non-uniformity in the main beam, tower, and anchorage zone. , and All are normalized according to preset rules; and These are the upper and lower limits of the cable tension; This refers to the allowable deviation for linearity; The target of Sota is off-center; This refers to the allowable deviation of the cable tower offset. and The allowable stress range for the component; This is the stress utilization factor; This refers to the jack's capacity limit. These are the normalized weighting coefficients for each indicator.
[0012] In one implementation, in step S5, the content stored in the tensioning decision knowledge base includes at least one of the following: baseline state of each construction stage, design target of the completed bridge, allowable range of cable force, control threshold of alignment and tower deviation, stress limit of components, jack calibration curve, single-stage tensioning increment limit, symmetrical tensioning construction rules, threshold of linkage between adjacent cables, geometric nonlinearity calculation rules, residual over-limit anomaly handling strategy and finite element model calling control strategy.
[0013] In one embodiment, step S5 further includes: When the predicted structural response of a candidate scheme is compared with the predicted structural observation response obtained by the finite element method, if the relative deviation exceeds a preset threshold, the linear cable force influence matrix is determined to be invalid. Sample points are then extracted within the candidate tension increment range, the finite element model is called to calculate the response of the sample points, and a second-order response surface model is established to replace the linear cable force influence matrix.
[0014] The present invention also provides an agent-based cable tensioning decision system for executing the above-described cable tensioning decision method, comprising: The data access module is used for basic data entry, including finite element models, construction stage data, tensioning equipment parameters, and field data. The physical-digital association module is used to build the model entity table and the field data channel table, and generate the model-field mapping set and mapping matrix; The tensioning decision knowledge base module is used to store the baseline state of the construction stage, the design target state, the threshold rules, the equipment calibration curves, the historical tensioning schemes, the abnormal handling rules, and the finite element tool calling strategies. The intelligent agent module integrates data processing, model correction, solution generation, solution verification, and execution feedback functions. The data area module is used to store model entity tables, field state vectors, mapping matrices, residual data, corrected model parameters, candidate tensioning schemes, verification results, and execution records.
[0015] The beneficial effects of this invention are as follows: Explicit associations are established between the finite element model entity and field measuring points, tensioning equipment, and construction stages, enabling field data to be projected onto the corresponding nodes, elements, and degrees of freedom of the finite element model according to mapping relationships. Data processing functions within the intelligent body are used for quality checks, spatiotemporal alignment, and temperature correction, reducing erroneous tensioning decisions caused by heterogeneous data sources, asynchronous sampling, or abnormal measuring points. The model correction function within the intelligent body corrects the finite element model during construction stages based on field data, ensuring that the tensioning scheme is based on the current actual structural state rather than a static design model. The scheme generation function within the intelligent body generates candidate tensioning schemes using cable force influence matrices or finite element response surfaces, reducing the cost of repeated manual calculations. The scheme verification function within the intelligent body recalculates candidate tensioning schemes at each level, simultaneously constraining cable force, alignment, tower offset, stress, adjacent cable influence, and equipment capacity. The execution feedback function within the intelligent body converts the tensioning scheme into field instructions and receives retest feedback, correcting subsequent schemes after each level of tensioning, reducing cable adjustment risks and improving the accuracy of bridge completion status control. Attached Figure Description
[0016] The invention will now be described in more detail with reference to embodiments and the accompanying drawings. Figure 1 A flowchart of the present invention is shown; Figure 2 This diagram illustrates the intelligent agent module, tension decision knowledge base, and finite element model calling relationship of the present invention. Figure 3 This diagram illustrates the process of generating, verifying, and correcting tensioning schemes in accordance with the present invention. Figure 4 The diagram shows the cable system arrangement, finite element model objects, and model-site mapping relationship of a certain bridge. Figure 5 This diagram shows a comparison of the target cable forces during the initial tensioning and secondary cable adjustment of a certain bridge. In the accompanying drawings, the same parts use the same reference numerals. The drawings are not to scale. Detailed Implementation
[0017] The invention will now be further described with reference to the accompanying drawings.
[0018] This invention provides an agent-based cable tension decision-making method, such as... Figures 1 to 3 As shown, it includes the following steps: Step S1: Input basic data into the database. Obtain the finite element model of the target cable-bearing structure, the target bridge design, the construction stage division, and the tensioning equipment parameters. Extract cable elements, main beam nodes, tower nodes, anchorage zone elements, temporary supports, construction loads, and boundary conditions from the finite element model to form a model entity table containing model entities, spatial locations, component types, construction stages, and mechanical degrees of freedom.
[0019] Step S2: Obtain the field data channel table and establish the relationship between the field data and the model entity table; Specifically, in step S2, the association relationship includes a model-site mapping set. Each mapping record in the set includes at least the finite element model node number or element number, cable number, beam end or tower end anchor point number, measuring point channel number, physical quantity type, coordinate transformation parameters, data conversion coefficient, reliability weight, construction stage number, and effective time window.
[0020] Step S3: Intelligent agent data processing. The intelligent agent reads the field data and performs channel integrity checks, abnormal transition checks, timestamp alignment, temperature compensation, equipment calibration coefficient conversion, and construction stage matching on the field data to obtain the field state vector of the current construction stage. The field state vector is represented as follows: ; in, For on-site cable tension, For hydraulic jacks, For the jack stroke, The displacement of the main beam. For Sota's deviation, For component stress, To measure the strain at the point, For the temperature field, For environmental load, This is a status indicator for the construction phase.
[0021] Step S4: Intelligent agent model correction. The intelligent agent projects the field state vector to the corresponding node, element or degree of freedom of the finite element model according to the relationship between the field data and the model entity table, calculates the residual between the field state vector and the finite element predicted state, and corrects the model parameters of the finite element model based on the residual to obtain the corrected finite element model for the current construction stage. Specifically, in step S4, the parameter sensitivity screening method is as follows: The parameters of each candidate model are perturbed, and the normalized comprehensive sensitivity coefficient of each parameter to cable force, main beam alignment, tower deviation and component stress is calculated by the central difference method. Low sensitivity parameters with comprehensive sensitivity coefficients lower than the preset threshold are fixed as the initial design values and do not participate in the optimization iteration. Define the residual vector for the construction phase as: ; The finite element model parameters are then corrected using the following objective function: ; in, Let be the residual vector for the t-th construction stage. For the model parameters to be corrected, These are the initial model parameters. For the finite element model of construction phase t, For finite element method or response surface method functions, This is the mapping matrix from the finite element calculation output to the field observations. The observation weight matrix, The regularization coefficient is . The cable force load vector, For construction load vectors, For the structural temperature field; The model parameters to be corrected include at least one of the following: effective elastic modulus of cable, equivalent stiffness correction factor for cable sag, anchorage retraction, temporary support stiffness, construction load deviation, temperature correction factor, weight correction factor for main beam segment, boundary condition correction factor, and elastic modulus correction factor for main beam / tower material.
[0022] Step S5: The agent generates candidate tensioning schemes. The agent calls the modified finite element model to apply unit perturbation calculations to the candidate tensioning cables, candidate tensioning sequence and candidate tensioning increment, forming a linear cable force influence matrix or finite element response surface. Then, it calls the tensioning decision knowledge base and uses the completed bridge cable force, main beam alignment, tower deflection, component stress, construction safety and equipment capacity as constraints to generate multiple sets of candidate tensioning schemes through multi-objective optimization. Specifically, in step S5, the linear relationship of the linear cable force influence matrix is as follows: ; The second-order response surface is represented as:
[0023] in, For candidate tension increment vectors, The structural response change vector, Let i be the predicted value of the i-th observed target. Let i be the baseline value for the i-th observed target in the current state. Let be the first-order influence coefficient of the tension increment of the k-th cable on the i-th observed target. Let be the second-order coupling influence coefficient of the tension increments of the k-th and l-th cables on the i-th observed target; The observation targets include at least one of the following: cable force, main beam alignment, tower deflection, component stress, temporary support reaction force, and anchorage zone stress. The multi-objective optimization function is: ; The constraints include: ; ; ; ,or ; ; ; in, To design the target state; Finite element prediction of candidate schemes; , This is a diagonal weight matrix used for normalizing physical quantities of different dimensions; For tensioning sequence risk indicators; For the number of adjacent cables that need to be adjusted synchronously, It is an index of strain energy or bending moment non-uniformity in the main beam, tower, and anchorage zone. , and All are normalized according to preset rules; and These are the upper and lower limits of the cable tension; This refers to the allowable deviation for linearity; The target of Sota is off-center; This refers to the allowable deviation of the cable tower offset. and The allowable stress range for the component; This is the stress utilization factor; This refers to the jack's capacity limit. These are the normalized weighting coefficients for each indicator; The content stored in the tensioning decision knowledge base includes at least one of the following: the baseline state of each construction stage, the design target of the completed bridge, the allowable range of cable force, the control threshold of alignment and tower deviation, the stress limit of components, the calibration curve of jacks, the limit of single-stage tensioning increment, the construction rules of symmetrical tensioning, the threshold of adjacent cable linkage, the calculation rules of geometric nonlinearity, the handling strategy for residual over-limit anomalies, and the finite element model calling control strategy. Furthermore, in step S5, when the predicted structural response value output by the linear cable force influence matrix for a candidate scheme is compared with the predicted structural observation response value obtained by the finite element solution, if the relative deviation exceeds a preset threshold, the linear cable force influence matrix is determined to be invalid. Sample points are then extracted within the candidate tension increment range, the finite element model is called to calculate the response of the sample points, and a second-order response surface model is established to replace the linear cable force influence matrix.
[0024] Step S6: Step-by-step finite element scheme verification. The agent performs finite element recalculation on each candidate tensioning scheme to determine whether the cable force deviation, main beam alignment deviation, tower displacement, component stress, anchorage zone stress, adjacent cable influence, jack capacity, and single-stage tensioning increment meet the preset constraints, and outputs the tensioning scheme that passes the verification. In particular, during the recalculation, if any of the following situations occur in any recalculation step: cable force exceeds the allowable range, tower deviation exceeds the warning threshold, main beam alignment deviation exceeds the limit, local component stress exceeds the limit, adjacent cable force change exceeds the linkage threshold, tensioning equipment capacity is insufficient, or construction steps are inconsistent with the site conditions, the candidate scheme will be rejected.
[0025] Step S7: The intelligent agent outputs on-site execution instructions, converting the verified tensioning scheme into on-site execution instructions, including tension cable number, tensioning sequence, target cable force, tensioning level, target increment for each level, jack hydraulic pressure or stroke control, pressure stabilization time, retest items, tensioning stop threshold, and abnormal handling strategy.
[0026] Step S8: After tensioning each level or each cable, recollect the field data and return to step S3 to revise the subsequent tensioning scheme.
[0027] It should be noted that the intelligent agent module can be implemented using rule-based reasoning, optimization algorithms, machine learning models, or a combination thereof, but its output must be subject to finite element recalculation and on-site constraint verification. The finite element model can be a full-bridge fine model, a construction stage model, a local anchorage zone model, a reduced-order model, or a response surface model. On-site cable forces can be obtained from vibration methods, pressure sensors, anchor sensors, intelligent steel strands, fiber optic grating sensors, or manually measured data. That is, this method uses the finite element construction stage model as the basis for mechanical calculation, on-site multi-source data as state feedback, and the internal functional arrangement of the intelligent agent as the decision-making mechanism to achieve automated processing of monitoring data, real-time correction of the finite element model, intelligent generation of tensioning schemes, hierarchical verification, command issuance, and closed-loop iterative updates, thereby improving the accuracy and intelligence level of tensioning construction control for cable-stayed bridges.
[0028] The present invention also provides an agent-based cable tensioning decision-making system, comprising: The data access module is used for basic data entry, including finite element models, construction stage data, tensioning equipment parameters, and field data. The physical-digital association module is used to build the model entity table and the field data channel table, and generate the model-field mapping set and mapping matrix; The tensioning decision knowledge base module is used to store the baseline state of the construction stage, the design target state, the threshold rules, the equipment calibration curves, the historical tensioning schemes, the abnormal handling rules, and the finite element tool calling strategies. The intelligent agent module integrates data processing, model correction, solution generation, solution verification, and execution feedback functions. The data area module is used to store model entity tables, field state vectors, mapping matrices, residual data, corrected model parameters, candidate tensioning schemes, verification results, and execution records. Among them, the integrated data processing function of the intelligent agent module is used to complete on-site data quality inspection, spatiotemporal alignment, temperature compensation and physical quantity conversion; the model correction function of the intelligent agent module is used to correct the finite element construction stage model based on on-site data and generate cable force influence matrix or finite element response surface; the scheme generation function of the intelligent agent module is used to generate candidate tensioning sequence, target cable force and tensioning increment; the scheme verification function of the intelligent agent module is used to verify the structural safety, equipment capacity and construction constraints of candidate tensioning schemes; and the execution feedback function of the intelligent agent module is used to output on-site execution instructions and receive retest data after tensioning.
[0029] In one embodiment, taking a large bridge as an example, the application of the present invention in the tensioning and adjustment process of cable stays is illustrated. This bridge is a single-tower, double-cable-stayed bridge with a span arrangement of 45+185+238+45=513m. It adopts a steel beam and steel tower structure, with the tower and beam rigidly connected. The entire bridge has a total of 4×17=68 cable stays, using a single-sided tensioning method at the beam ends. The longest cable is approximately 282.5m, and the shortest is approximately 123.5m. The cables are made of monofilament epoxy-coated prestressed steel strands with a nominal diameter of 15.2mm, a tensile strength of 1860MPa, and an elastic modulus of 1.95×10⁻⁶. 5 The maximum strength is 15.2-85 MPa, the minimum strength is 15.2-31 MPa, the maximum weight of a single cable is approximately 29 tons, the cable spacing on the beam is approximately 9 meters, and the standard cable spacing on the tower is approximately 2.5 meters. Both the beam and tower ends are anchored with steel anchor beams. Due to the asymmetry between the main span and side spans, the spatially twisted irregular steel towers, and the construction phases involving dismantling and lowering the beams, the interaction between the main beam deflection, the main tower offset, and the cable tension is significant. The cable system layout, finite element model objects, and model-site mapping relationship of this bridge are as follows: Figure 4 As shown, the process for generating the tensioning scheme and subsequent cable adjustment using the method of this invention is as follows: The overall layout drawing of a bridge, cable system layout drawing, full-bridge finite element model, stay cable calculation model, tensioning sequence scheme, initial tensioning target cable force table, adjustment cable target cable force table, and on-site re-measurement records are written into the data area to form a model entity table, a site access table, and a construction stage table. The model entity table includes at least the main girder nodes, bridge tower nodes, cable elements, cross brace elements, anchorage zone elements, temporary supports, and construction loads. The site access table includes at least cable force, jack hydraulic pressure, jack stroke, main girder alignment, main tower offset, stress, temperature, and construction stage status. In other words, the engineering data is no longer merely saved as construction background, but becomes the calculation input for subsequent model correction and tensioning scheme generation. The data area is divided into model data, field data, mapping data, residual data, knowledge base data, scheme data, verification data, and execution data according to the engineering purpose; Model data storage includes finite element model paths, model entity tables, construction phase model snapshots, and response surfaces; On-site data storage includes original remeasured values, quality inspection results, temperature correction results, and construction stage identifiers; Mapped data storage model - on-site mapping set R, mapping matrix C, reliability weights, and effective time window; The scheme data, verification data, and execution data respectively store candidate tensioning schemes, finite element recalculation conclusions, reasons for rejection, on-site task sheets, retest data, and process record numbers; The four cable types (SA, SB, MA, MB) are mapped to cable elements in the finite element model, and their beam-end or tower-end anchorage points are associated with the tensioning equipment number, measurement channel number, and construction stage number. For main beam alignment measurement points, the measured coordinates are transformed to the finite element model coordinate system; for main tower deviation measurement points, they are associated with the nodal degrees of freedom at the corresponding height of the tower; for stress measurement points, they are associated with the corresponding element output positions of the main beam, tower, cross brace, or anchorage zone. When the remeasured data enters the system after the scaffolding is dismantled and the beam is lowered, the data processing function within the intelligent body first determines whether the cable force, elevation, deviation, and stress data are consistent with the current construction stage. The data is matched, and then the model correction function calculates the residual between the measured state and the finite element prediction state. After the field data enters the data area, the data processing function in the intelligent body performs at least the following processing: checking whether there are sudden jumps, missing measurements, or exceeding the physical upper limit of the cable force; checking whether the hydraulic pressure and stroke of the jack are consistent with the equipment calibration curve; converting the total station or leveling measurement results to the finite element model coordinate system; forming a temperature field by dividing the temperature measurement points into component zones, and correcting the temperature influence in cable force, strain, and alignment; converting construction records such as cable hanging completion, frame dismantling and beam lowering, second-phase dead load loading, and temporary cable replacement into construction stage identifiers; and generating a field state vector. And write it into the corresponding stage; The system reads the existing candidate sequences in the tensioning decision knowledge base and selects the following as candidate sequences: tensioning from the far end of the beam to the near end of the tower, tensioning from the near end of the tower to the far end, tensioning from the middle cable outwards, tensioning odd-numbered cables first and then even-numbered cables, and tensioning cables after dismantling the frame in stages. The scheme generation function in the intelligent body uses structural stress, main beam alignment, main tower deviation, cable force deviation, construction difficulty, safety risk, and process continuity as evaluation indicators and calls the modified finite element model for comparison. For this bridge, tensioning from the near end of the tower symmetrically to the far end can be regarded as one of the preferred candidate sequences. This sequence is not a simple reference to human experience, but is written into the knowledge base as a candidate sequence and then sorted by cable force influence matrix and finite element recalculation. Specifically, the selection of tensioning sequence can be compiled into the following engineering rule table:
[0030] The scheme generation function within the intelligent system is based on the current construction stage model and the initial tension target cable force table. It applies a unit tension increment to each candidate cable and calculates the cable force influence matrix. If the prediction error of the linear influence matrix for the main beam alignment or main tower deviation exceeds the threshold, sample points are extracted within the allowable tension increment range to establish a response surface model. Based on the target cable force, main beam alignment, main tower deviation, stress limit, tensioning equipment capacity, and single-level increment limit, the system outputs the initial tensioning sequence, the target cable force for each cable, the graded tensioning values, and the retest items. The scheme verification function within the intelligent system then recalculates step by step according to the tensioning level. Only after confirming that the cable force, alignment, deviation, stress, and equipment capacity all meet the constraints are the candidate schemes listed as executable schemes. The candidate tensioning scheme can be represented as ,in, i represents the tensioning order. For the increase in tension of each cable, For graded tensioning stages, For voltage stabilization time, For the retest project, when the current stage shows that the main beam alignment on the main span side is too low, the tower is offset towards the main span side, and the cable force of some main span side cables is insufficient, the scheme generation function in the intelligent body does not directly assign all the deviations to the single cable with the largest deviation. Instead, it identifies cable combinations that can improve both the main beam alignment and the tower offset based on the cable force influence matrix, generates a graded tensioning scheme, and limits the single-level increment of a single cable to avoid excessive passive changes in adjacent cables. The initial tension target cable force is not a fixed final value, but rather a stage target under the current construction boundary and current model state. Taking cable number examples, the initial tension target cable force for SA17 is approximately 3735kN, for SB17 it is approximately 2780kN, and for MB17 it is approximately 1070kN; the initial tension target cable force for MA7, which is closer to the middle, is approximately 4510kN, and for MB7 it is approximately 2345kN. After these target values are written into the scheme data area, the scheme generation function in the intelligent body converts the hydraulic pressure or stroke control amount according to the single-stage tension increment limit and the equipment calibration curve. The scheme verification function in the intelligent body then recalculates the main beam alignment, main tower offset, and changes in adjacent cable forces for the corresponding stage. Specifically, the results of the step-by-step verification of the candidate solutions can be written into the verification data as shown in the table below:
[0031] After the scaffolding is dismantled and the beams are lowered, the model is corrected. Following this, the system reads the measured data of cable forces, main beam alignment, main tower displacement, and stress for the entire bridge and compares them with the predicted values from the finite element construction phase. If the residual values of some cable forces, main beam elevations, or main tower displacements exceed a threshold, the model correction function within the intelligent body first performs parameter sensitivity screening, then corrects parameters such as cable equivalent stiffness, anchorage retraction, temporary support stiffness, segment weight, or boundary conditions to obtain a corrected finite element model for secondary cable adjustments. This step transforms the construction boundary change of dismantling and lowering the beams into model parameters and conditions. The system updates the state residuals, rather than just noting the deviations in the field records. For example, when the measured cable force of multiple adjacent cables is low as a whole and the main beam elevation is low, the model correction function prioritizes identifying segment weight deviations or temporary support stiffness deviations. When the measured increment of a single cable is lower than the target increment and the hydraulic pressure of the jack reaches the target value, the model correction function prioritizes identifying anchor retraction, friction loss, or cable length correction. If additional tensioning would lead to insufficient equipment capacity or excessive structural response, the system marks the candidate scheme as requiring manual review and reallocates the tensioning increment of subsequent cables. The scheme generation function within the intelligent system generates a secondary cable adjustment scheme based on the modified finite element model and the cable adjustment target force table. The cable adjustment scheme includes at least the cable adjustment sequence, cable number, target cable force, single-stage tension increment, jack hydraulic pressure or stroke control, pressure stabilization time, retest items, tension stop threshold, and manual verification requirements. The scheme verification function within the intelligent system recalculates the cable adjustment process step by step. If there is a reverse over-adjustment of the main beam alignment, the main tower deviation exceeding the limit, excessive changes in adjacent cable forces, or insufficient equipment capacity, the candidate scheme is rejected, and the cable adjustment increment is reallocated by the scheme generation function within the intelligent system. The cable adjustment scheme that passes the verification is converted into a field task order. After the first-stage cable adjustment or the adjustment of one cable is completed on-site, the retest data continues to be written into the record for subsequent model correction and scheme recalculation. The target cable force during the secondary cable adjustment stage reflects the model correction results after the dismantling of the girder, changes in the second-phase dead load, and temporary boundary conditions. For example, SA17 can be adjusted from the initial tension target of approximately 3735 kN to the secondary cable adjustment target of approximately 3495 kN; SB17 can be adjusted from approximately 2780 kN to approximately 2240 kN; MA17 achieves a target value of approximately 2245 kN during the cable replacement or adjustment stage; and MB17 can be adjusted from approximately 1070 kN to approximately 3510 kN. Similarly, MA8 can be adjusted from the initial tension target of approximately 3985 kN to approximately 4650 kN, and MB8 can be adjusted from approximately 2250 kN to approximately 3265 kN. This difference demonstrates that cable adjustment is not a simple replication of the initial tension target, but rather a redistribution of the cable force increment based on the overall bridge cable force, main girder alignment, main tower offset, and measured stress conditions after dismantling the girder. A comparison of the target cable forces between the initial tensioning and the secondary cable adjustment is provided. Figure 5 As shown; Through the above processing, the tensioning and cable adjustment process of a certain bridge can compare the finite element model, on-site cable force, main beam alignment, main tower deviation, stress data and tensioning equipment parameters under the same construction stage, so that the subsequent cable adjustment scheme no longer depends solely on the offline target value, but is corrected based on the actual measured state after the dismantling of the frame and the lowering of the beam. In the description of this invention, it should be understood that the terms "upper", "lower", "bottom", "top", "front", "rear", "inner", "outer", "left", "right", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.
[0032] While the invention has been described herein with reference to specific embodiments, it should be understood that these embodiments are merely examples of the principles and applications of the invention. Therefore, it should be understood that many modifications can be made to the exemplary embodiments, and other arrangements can be designed without departing from the spirit and scope of the invention as defined by the appended claims. It should be understood that different dependent claims and features described herein can be combined in ways different from those described in the original claims. It is also understood that features described in conjunction with individual embodiments can be used in other described embodiments.
Claims
1. A cable tensioning decision-making method based on intelligent agents, characterized in that, Includes the following steps: Step S1: Input basic data into the database, obtain the finite element model of the target cable-bearing structure, the target bridge design, the construction stage division and tensioning equipment parameters, extract cable elements, main beam nodes, tower nodes, anchorage zone elements, temporary supports, construction loads and boundary conditions from the finite element model, and form a model entity table containing model entities, spatial location, component type, construction stage and mechanical degrees of freedom; Step S2: Obtain the field data channel table and establish the relationship between the field data and the model entity table; Step S3: Intelligent agent data processing. The intelligent agent reads the field data and performs channel integrity checks, abnormal transition checks, timestamp alignment, temperature compensation, equipment calibration coefficient conversion, and construction stage matching on the field data to obtain the field state vector of the current construction stage. Step S4: Intelligent agent model correction. The intelligent agent projects the field state vector to the corresponding node, element or degree of freedom of the finite element model according to the relationship between the field data and the model entity table, calculates the residual between the field state vector and the finite element predicted state, and corrects the model parameters of the finite element model based on the residual to obtain the corrected finite element model for the current construction stage. Step S5: The agent generates candidate tensioning schemes. The agent calls the modified finite element model to apply unit perturbation calculations to the candidate tensioning cables, candidate tensioning sequence and candidate tensioning increment, forming a linear cable force influence matrix or finite element response surface. Then, it calls the tensioning decision knowledge base and uses the completed bridge cable force, main beam alignment, tower deflection, component stress, construction safety and equipment capacity as constraints to generate multiple sets of candidate tensioning schemes through multi-objective optimization. Step S6: Step-by-step finite element scheme verification. The agent performs finite element recalculation on each candidate tensioning scheme to determine whether the cable force deviation, main beam alignment deviation, tower displacement, component stress, anchorage zone stress, adjacent cable influence, jack capacity, and single-stage tensioning increment meet the preset constraints, and outputs the tensioning scheme that passes the verification. Step S7: The intelligent agent outputs on-site execution instructions; Step S8: After tensioning each level or each cable, recollect the field data and return to step S3 to revise the subsequent tensioning scheme.
2. The agent-based cable tensioning decision-making method according to claim 1, characterized in that, In step S2, the association relationship includes a model-site mapping set. Each mapping record in the set includes at least the finite element model node number or element number, cable number, beam end or tower end anchor point number, measuring point channel number, physical quantity type, coordinate transformation parameters, data conversion coefficient, reliability weight, construction stage number, and effective time window.
3. The agent-based cable tension decision-making method according to claim 2, characterized in that, In step S3, the field state vector is represented as: ； in, For on-site cable tension, For hydraulic jacks, For the jack stroke, The displacement of the main beam. For Sota's deviation, For component stress, To measure the strain at the point, For temperature field, For environmental load, This is a status indicator for the construction phase.
4. The agent-based cable tensioning decision-making method according to claim 3, characterized in that, In step S4, the parameter sensitivity screening method is as follows: The parameters of each candidate model are perturbed, and the normalized comprehensive sensitivity coefficient of each parameter to cable force, main beam alignment, tower deviation and component stress is calculated by the central difference method. Low sensitivity parameters with comprehensive sensitivity coefficients lower than the preset threshold are fixed as the initial design values and do not participate in the optimization iteration. Define the residual vector for the construction phase as: ； The finite element model parameters are then corrected using the following objective function: ； in, Let be the residual vector for the t-th construction stage. For the model parameters to be corrected, These are the initial model parameters. For the finite element model of construction phase t, For finite element method or response surface method functions, This is the mapping matrix from the finite element calculation output to the field observations. The observation weight matrix, The regularization coefficient is . The cable force load vector, For construction load vectors, This represents the temperature field of the structure.
5. The agent-based cable tensioning decision-making method according to claim 4, characterized in that, In step S4, the model parameters to be corrected include at least one of the following: effective elastic modulus of cable, equivalent stiffness correction coefficient for cable sag, anchor retraction, temporary support stiffness, construction load deviation, temperature correction coefficient, weight correction coefficient for main beam segment, boundary condition correction coefficient, and elastic modulus correction coefficient for main beam / tower material.
6. The agent-based cable tensioning decision-making method according to claim 5, characterized in that, In step S5, the linear relationship of the linear cable force influence matrix is as follows: ； The second-order response surface is represented as: in, For candidate tension increment vectors, The structural response change vector. Let i be the predicted value of the i-th observed target. Let i be the baseline value for the i-th observed target in the current state. Let be the first-order influence coefficient of the tension increment of the k-th cable on the i-th observed target. Let be the second-order coupling influence coefficient of the tension increment of the k-th and l-th cables on the i-th observed target; The observation targets include at least one of the following: cable force, main beam alignment, tower deviation, component stress, temporary support reaction force, and anchorage zone stress.
7. The agent-based cable tensioning decision-making method according to claim 6, characterized in that, In step S5, the multi-objective optimization function is: ； The constraints include: ； ； ； ,or ; ； ； in, To design the target state; Finite element prediction of candidate schemes; , This is a diagonal weight matrix used for normalizing physical quantities of different dimensions; For tensioning sequence risk indicators; For the number of adjacent cables that need to be adjusted synchronously, It is an index of strain energy or bending moment non-uniformity in the main beam, tower, and anchorage zone. , and All are normalized according to preset rules; and These are the upper and lower limits of the cable tension; This refers to the allowable deviation for linearity; The target of Sota is off-center; This refers to the allowable deviation of the cable tower offset. and The allowable stress range for the component; Stress utilization coefficient; This refers to the jack's capacity limit. These are the normalized weighting coefficients for each indicator.
8. The agent-based cable tensioning decision-making method according to claim 7, characterized in that, In step S5, the content stored in the tensioning decision knowledge base includes at least one of the following: the baseline state of each construction stage, the design target of the completed bridge, the allowable range of cable force, the control threshold of alignment and tower deviation, the stress limit of components, the calibration curve of the jack, the limit of single-stage tensioning increment, the symmetrical tensioning construction rules, the threshold of adjacent cable linkage, the geometric nonlinear calculation rules, the residual over-limit anomaly handling strategy, and the finite element model calling control strategy.
9. The agent-based cable tensioning decision-making method according to claim 8, characterized in that, Step S5 also includes: When the predicted structural response of a candidate scheme is compared with the predicted structural observation response obtained by the finite element method, if the relative deviation exceeds a preset threshold, the linear cable force influence matrix is determined to be invalid. Sample points are then extracted within the candidate tension increment range, the finite element model is called to calculate the response of the sample points, and a second-order response surface model is established to replace the linear cable force influence matrix.
10. A cable tensioning decision-making system based on an agent, used to execute the cable tensioning decision-making method based on an agent as described in any one of claims 1 to 9, characterized in that, include: The data access module is used for basic data entry, including finite element models, construction stage data, tensioning equipment parameters, and field data. The physical-digital association module is used to construct the model entity table and the field data channel table, and generate the model-field mapping set and mapping matrix. The tensioning decision knowledge base module is used to store the baseline state of the construction stage, the design target state, the threshold rules, the equipment calibration curves, the historical tensioning schemes, the abnormal handling rules, and the finite element tool calling strategies. The intelligent agent module integrates data processing, model correction, solution generation, solution verification, and execution feedback functions. The data area module is used to store model entity tables, field state vectors, mapping matrices, residual data, corrected model parameters, candidate tensioning schemes, verification results, and execution records.