Bridge intelligent analysis method and system based on bridge engineering deduction model

Abstract

The application discloses a bridge intelligent analysis method and system based on a bridge engineering deduction model, and constructs a bridge engineering deduction model composed of physical, learning, symbolic and state evolution submodels: the physical submodel accurately predicts, the learning submodel second-level approximates and labels uncertainty, the symbolic submodel provides norm constraint and traceability, and the state evolution submodel describes state transition with the physical anchor points of unstressed length and unstressed curvature; unified interface encapsulates five types of operations of prediction, parameter change deduction, scheme deduction, conclusion verification and actual measurement calibration, and digital non-reproducible is intercepted; an automatic processing unit is sequentially arranged and executed according to a dependency graph, and is explored by a learning submodel and reviewed by a physical submodel; and finally, a conclusion object with traceability and margin and reviewed by a human being is output. The application makes the bridge engineering artificial intelligence conclusion reproducible, traceable and signable.

Description

Technical Field

[0001] This invention belongs to the interdisciplinary field of bridge engineering calculation and artificial intelligence. Specifically, it relates to a method for constructing a bridge engineering simulation model, as well as a bridge intelligent analysis method, system, electronic device and computer-readable storage medium based on the model, which can be applied to bridge design, construction control and old bridge evaluation. Background Technology

[0002] Existing bridge analysis software (such as Midas Civil, Bridge Doctor, OSIS, etc.) relies on engineers to manually drive modeling, analysis, calculation, and report generation. While it boasts powerful computational capabilities, all conclusions are judged by humans, failing to provide physical verification for AI-generated engineering recommendations. As large language models enter engineering design assistance scenarios, AI may generate multi-step engineering recommendations, but these generally suffer from issues such as disordered steps, missing steps, and false values. The conclusions lack physical and regulatory basis, and are unreproducible, untraceable, and uncertifiable.

[0003] The universal world model uses machine learning to obtain an internal predictive model for automated unit planning and decision-making. However, its predictions do not guarantee compliance with structural mechanics conservation laws, and the results drift with the solution path, making them unsuitable for direct engineering decision-making. Digital twins mirror physical entities using Building Information Modeling (BIM) or monitoring data, primarily for display and early warning, but lack the ability to extrapolate parameter changes and perform scenario simulations. Neither approach solves the unique technical problem in bridge engineering: the conclusions of engineering AI must be reproducible, traceable, and verifiable.

[0004] In the field of bridge construction simulation, the traditional approach is "unit birth and death + state accumulation": new units are activated step by step according to the construction sequence, and the deformation state of the previous stage is accumulated as the initial condition for the next stage. This method assembles new units on deformed nodes, assuming that the new units inherit the deformation of the current node and record non-existent initial strains (false initial strains). This leads to inconsistencies between the phased construction results and the results of a single completed bridge, and the errors accumulate along the construction sequence, making it difficult to pinpoint specific stages and units. There is a publicly available theory in the bridge field regarding stress-free state control: when the stress-free length and stress-free curvature of each unit are determined, the internal forces and alignment of the completed bridge are independent of the construction process. However, this theory has not yet been implemented in general software or integrated with artificial intelligence decision-making loops.

[0005] Therefore, existing technologies lack a bridge engineering simulation model and intelligent analysis method that enables the conclusions of engineering artificial intelligence to be reproducible, traceable, and verifiable, and ensures that the construction simulation results are guaranteed by the structure rather than by manual sampling. Summary of the Invention

[0006] Purpose of the Invention: To address the aforementioned shortcomings of existing technologies, this invention provides a bridge intelligent analysis method and system based on a bridge engineering simulation model. This method ensures that all predictions, simulations, plans, and conclusions of the automated processing unit are executed or verified through an engineering simulation model that incorporates "physical constraints + normative constraints + reproducible verification," thereby making the conclusions of engineering artificial intelligence reproducible, traceable, and verifiable. The bridge engineering simulation model referred to in this invention is composed of a physical sub-model, a learning sub-model, a symbolic sub-model, and a state evolution sub-model, and provides five types of operations—prediction, parameter change simulation, scheme simulation, conclusion verification, and experimental calibration—through a unified interface. Its difference from the general world model lies in the fact that all outputs can be reproduced and verified through a deterministic computation kernel.

[0007] The technical solution adopted in this invention is: a bridge intelligent analysis method based on a bridge engineering simulation model, comprising the following steps: Step 1: Construct a bridge engineering simulation model: register a physical sub-model, a learning sub-model, a symbolic sub-model, and a state evolution sub-model; wherein, the physical sub-model provides accurate response prediction based on a deterministic numerical calculation kernel; the learning sub-model provides approximate prediction with uncertainty based on a surrogate model; the symbolic sub-model provides constraints and their sources based on a knowledge base of standard provisions; and the state evolution sub-model describes the state transitions of the bridge throughout its entire life cycle based on the construction stage sequence. Step 2: Provide a unified interface: encapsulate five types of operations: prediction, parameter change deduction, scheme deduction, conclusion verification, and actual measurement calibration. This ensures that the predictions, deductions, and conclusions of the automated processing unit and the user interface are all executed or verified through this unified interface, avoiding bypass channels that circumvent physical and regulatory constraints. Step 3, Ordered Execution: The automated processing unit selects the workflow through the function registry, generates the execution sequence by sorting the dependency graph topology, and executes it step by step by the ordered executor. Before each step is executed, a precondition check is performed, and after each step is executed, a postcondition check is performed. In the exploration phase, the learning sub-model is forced to be used and uncertainties are labeled with surrogate estimation. In the issuance phase, the physical sub-model is forced to be used for precise verification, and the direct generation of conclusion objects from approximate results is prohibited. Step 4, Reproducibility Verification: Verify whether the values ​​output by the automated processing unit can be reproduced in the bridge engineering simulation model; the verification is based on the location-working condition-unit-component quadruple, checking the running identifier and model hash corresponding to the values. If they cannot be reproduced, they are blocked and not included in the conclusion object. Step 5, Issuance and Calibration: Generate a conclusion object containing the conclusion, supporting clauses, margin, operation identifier, model hash, and manual review status; the automated processing unit can only produce the status pending review, which takes effect after manual review; and the bridge engineering simulation model parameters are corrected in reverse using measured monitoring data and measured calibration operations, and then proceed to the next round of prediction.

[0008] Furthermore, the rules of the symbolic sub-model in step one carry a specification version field and clause location, and automatically identify expired rules when the specification version changes.

[0009] Furthermore, in step one, the state evolution sub-model uses the stress-free length and stress-free curvature of the unit as physical anchor points: the model is independently reconstructed and solved for each construction stage, and the newly installed units are assembled into the reference configuration according to the explicitly given stress-free state quantities; and the consistency between the results of one-time bridge completion and phased construction, and the consistency between the results of forward assembly and backward dismantling are used as closure criteria for automatic verification. Furthermore, in step one, an influence matrix is ​​constructed using the stress-free cable length as a design variable, and the stress-free cable length adjustment corresponding to the target bridge completion state is solved in reverse, outputting the material cutting length correction and the jack tensioning stroke.

[0010] Furthermore, in step two, the parameter change deduction is used to return the difference vector of the parameter change relative to the baseline response and its associated components; the scheme deduction is used to deduce the candidate action sequence into a state trajectory that evolves with the construction stage.

[0011] Furthermore, the entries in the function registry in step three include input / output contracts, preconditions, precision levels, and applicable bridge types; adding a new function means registering a line and making the automated processing unit automatically visible; the language model is only responsible for workflow selection and parameter suggestions, and has no right to change the execution order.

[0012] Furthermore, the conclusion object in step three includes the conclusion, the supporting clauses, the margin, the running identifier, the model hash, and the manual review status; the transition from the pending review status to the effective status requires manual action.

[0013] Furthermore, the bridge application domain includes a new bridge design domain and an old bridge reinforcement domain, with specifications, verification procedures, and deliverable templates configured to distinguish between different application domains.

[0014] A system employing the aforementioned intelligent bridge analysis method based on a bridge engineering simulation model includes a processor and a memory. The memory stores a computer program, and the processor executes the computer program to implement the aforementioned intelligent bridge analysis method.

[0015] The present invention has the following beneficial effects: (1) The conclusions of the engineering artificial intelligence are credible. Each value is verified by physical and standard verification, and can be reproduced, traced and issued; (2) The prediction of the engineering simulation model does not drift with the solution path. The stress-free physical anchor point ensures that the phased construction and the one-time bridge completion results are consistent by the structure. The error can be located to the specific stage and unit; (3) The exploration and issuance are graded. The exploration mode is screened in seconds and the issuance mode is accurately verified, achieving both efficiency and credibility; (4) The actual measurement calibration closed loop ensures that the engineering simulation model and the real bridge are continuously aligned; (5) The conclusion object has a responsible subject. The artificial verification state machine makes the artificial intelligence conclusions responsible. Attached Figure Description

[0016] Figure 1 is a schematic diagram of the four sub-models of the bridge engineering simulation model of the present invention; Figure 2 is a schematic diagram of the relationship between the unified interface and the automated processing unit / user call of the present invention; Figure 3 is a schematic diagram of the reproduction verification and conclusion issuance process of this invention; Figure 4 is a schematic diagram of the physical anchor points of the stress-free state quantity and the construction stage of the present invention; Figure 5 is a schematic diagram of the ordered execution process of the present invention. Detailed Implementation

[0017] The technical solution of the invention will be described in detail below with reference to the accompanying drawings. It should be understood that these embodiments are only used to illustrate the invention and not to limit the scope of the invention. After reading this invention, any modifications of the invention in various equivalent forms made by those skilled in the art fall within the scope defined by the appended claims.

[0018] As shown in Figure 1, the bridge engineering simulation model 101 of this invention consists of four sub-models: a physical sub-model 102, a learning sub-model 103, a symbolic sub-model 104, and a state evolution sub-model 105. It provides five types of operations to the automated processing unit 107 and the user / interface 108 via a unified interface layer 106. The four sub-models are: (1) Physical sub-model: Encapsulates the deterministic numerical calculation kernel, supports static, modal and moving load analysis of bridge types such as simply supported beams, continuous beams, cable-stayed bridges and suspension bridges; for the same input, the output is deterministic and can be run repeatedly, and the output records the run identifier run_id and model hash model_hash.

[0019] (2) Learning sub-model: Encapsulate the surrogate model (e.g., Gaussian process regression, neural network regression or radial basis function interpolation), train it with offline samples of the physical sub-model, provide second-level approximate predictions, and output the uncertainty of the surrogate estimate (e.g. prediction variance ±σ).

[0020] (3) Symbol sub-model: encapsulates the knowledge base of standard clauses and rule engine; each rule carries the standard version field and clause location (such as Clause 4.3.1 of "General Specification for Design of Highway Bridges and Culverts" JTG D60-2015), and the rule value can be traced back to the original clause; when the standard version changes, expired rules are automatically identified and marked.

[0021] (4) State evolution sub-model: encapsulates the construction stage sequence and unit state; each unit carries the stress-free length S0 and stress-free curvature κ0 as physical anchor points, and carries the reference configuration source identifier; the construction stage sequence defines the new unit set, demolition unit set, boundary change, tensioning event, time point and concrete age for each stage.

[0022] As shown in Figure 2, both the automated processing unit 201 and the user / interface 202 call the five types of operations encapsulated by the unified interface 203 of the engineering simulation model—prediction 204, parameter change simulation 205, scheme simulation 206, conclusion verification 207, and actual measurement calibration 208, and are connected to four sub-models 209: (1) Prediction: Input the current state and action, and output the response of the physical sub-model or the learning sub-model; (2) Parameter change deduction: Input parameter change, output change difference vector and its associated components relative to the baseline response; (3) Scheme deduction rollout: Input candidate action sequence, output the state trajectory that evolves with the construction stage; (4) Verify the conclusion: Input the numerical values ​​and their quadruple locations in the conclusion, rerun and reproduce them in the physical sub-model and compare them; (5) Actual measurement calibration sync: Input actual measurement monitoring data and reverse correct the parameters of the engineering simulation model.

[0023] The automated processing unit shares this interface with the user interface; all predictions, deductions and conclusions of the automated processing unit must be executed or verified through this interface, and there is no bypass.

[0024] Example: Intelligent design and construction control of a 400m main span double-tower cable-stayed bridge Input objects and initial conditions: span arrangement (main span 400m, double towers), bridge width, material parameters, main beam and tower section parameters, specifications of 80 stay cables (area A, elastic modulus E, self-weight w), target bridge alignment and cable force range, action sequence of 60 construction stages (including 80 cable tensioning events), and optional measured monitoring data.

[0025] Step 1 is implemented as follows: Register the physical sub-model (beam-cable geometric nonlinear finite element calculation kernel), the learning sub-model (Gaussian process agent, training samples are generated offline by the physical sub-model), the symbolic sub-model (JTG series specification rule base, rules with clause location and version fields), and the state evolution sub-model (60-stage sequence; beam elements carry S0 and κ0, cable elements carry S0). Each sub-model is registered in the registry with the following information: model identifier, type, input / output contract, accuracy profile (fast / exact), applicable bridge type, and state.

[0026] Step Two Implementation: Generate a unified interface instance for the four sub-models. Example Call: The automated processing unit requests `predict(current stage snapshot, next batch of cable tension)`, the exploration mode routes to the learning sub-model and labels it "surrogate estimate ±σ"; the issuance mode routes to the physical sub-model for exact solution. `counterfactual` returns "the difference in the impact of a 5% overall increase in this batch of cable tension on the main beam bending moment and alignment"; `rollout` returns the 60-stage state trajectory; `verify` performs reproducibility verification on each conclusion value; `sync` receives the measured cable tension sequence.

[0027] Step 3 Implementation: As shown in Figure 5, the workflow is determined by the user objective 501. The macro-planning (large language model) 502 selects the workflow and parameters. Based on the functional registry and dependency graph 503 topological sorting, the ordered executor 504 executes step by step under the wrapping of 507 precondition verification and 508 post-verification. The execution trajectory is recorded 505 and 506 Verdict is generated and issued. The functional registry registers the capabilities of model building, form finding, load arrangement, modal analysis, influence line solving, combination verification, report generation, etc. The entries include input and output contracts, preconditions, accuracy files and applicable bridge types. To achieve the user objective of "the flattest bridge alignment and uniform cable tension," the automated processing unit selects a workflow template from the registry. The system constructs a directed acyclic graph based on data dependencies; for example, impact coefficient calculation depends on modal analysis results, load combinations depend on live and dead load effects, and verification depends on combined effects. Kahn topology sorting is used to generate the execution sequence. The ordered executor executes "pre-verification → execution → post-verification" step-by-step, recording the execution trajectory at each step (step identifier, input summary, output summary, run identifier, timestamp). The language model only participates in workflow selection and parameter suggestions and has no authority to change the execution order.

[0028] Step 4 Implementation: As shown in Figure 3, after the value 301 output by the automated processing unit is extracted and verified by the engineering simulation model (303), it is not reproducible (304) and is intercepted and judged as a false value. If it is reproducible (305), it is written into the conclusion and forms a (306) Verdict (draft), which takes effect after manual review (307). For each value output by the automated processing unit, the (location-condition-unit-component) quadruple and its source running identifier run_id and model hash model_hash are extracted. The verify function is called to rerun the same input in the physical sub-model, and the rerun result is compared with the output value. If the relative deviation exceeds the issuance tolerance (e.g., 0.5%), it is judged as a false value, the value is intercepted, and the process is reverted to the automated processing unit step that generated the value. The verified result is archived along with the evidence package (rerun input, output, tolerance, signature).

[0029] Step 5 Implementation: As shown in Figure 3, a conclusion object, Verdict, is generated, containing six fields: conclusion, supporting clause (symbol sub-model location), margin, running identifier, model hash, and manual review status. The state machine is: draft (pending review, can be generated by the automated processing unit) → reviewed (manual review passed, effective) → archived. When rendering the conclusion, it is verified that run_id and model_hash can still be reproduced; otherwise, it is downgraded to "invalid". During the running period, the measured cable forces are updated using recursive least squares or Bayesian correction to update the physical sub-model parameters (e.g., equivalent stiffness of the cable, effective stiffness of the main beam) before entering the next round of prediction.

[0030] Stress-free state quantity physical anchor point As shown in Figure 4, the state evolution sub-model uses the stress-free length S0 and stress-free curvature κ0 as the first type of state variables, progressing according to construction stages (stage k-1 401, stage k 402, stage k+1 403). Each stage records a stage snapshot 404 (internal force / displacement / cable force / updated S0 / κ0), which is automatically verified by the closure criterion 405 and used for stress-free cable length control 406. The initial source of the stress-free state variables can be: explicit assignment of cutting length and precast alignment, or back-calculation from the current configuration and internal forces. , , where Le is the stress length of the element, N is the axial force, M is the bending moment, E is the elastic modulus, I is the moment of inertia of the section, and κ is the stress curvature.

[0031] For each construction stage k, the system independently reconstructs the solution model based on the currently installed unit set: the newly installed units are assembled according to their own explicit S0 / κ0 reference configurations, without inheriting the deformation of the current nodes, thereby eliminating spurious initial strains; after applying the loads and boundary conditions of this stage, the static equilibrium is solved independently; the results of each stage are recorded as repeatable snapshots (displacement, internal force, cable force, and updated stress-free state quantities).

[0032] Automatic verification of closure criteria: For the same set of stress-free state quantities, perform "one-time bridge completion" and "phased construction" solutions respectively, and compare the internal forces and alignment of the completed bridge; then perform "forward assembly" and "reverse dismantling" solutions, and compare the state of each stage; the tolerance is graded as follows: the relative deviation of internal force / alignment is less than 0.5%, and the relative deviation of cable force is less than 2%. If the tolerance is exceeded, the specific stage and unit are automatically located and reported. The time-dependent effect (shrinkage and creep) is expressed as the time-varying increment of stress-free force: ΔS0 = S0·(Δεcs+ Δεcr), Δκ0 = (Δεtop - Δεbot) / h, where Δεcs is the shrinkage strain increment, Δεcr is the creep strain increment, Δεtop and Δεbot are the fiber strain increments at the top and bottom of the section respectively, and h is the section height.

[0033] Stress-free cable length adjustment Using the stress-free cable length S0 of the 80 stay cables as the design variable x∈R^80, an influence matrix A∈R^(m×80) is constructed: a unit stress-free cable length disturbance ΔS0 is applied to the j-th cable, and the physical sub-model is called to solve the problem. The difference between the target responses (bridge alignment control points, main girder bending moment, and cable force uniformity index) is taken as the j-th column. The response difference between the target bridge state and the current state is... Solve the least squares problem with regularization:

[0034] Where W is the target weight diagonal matrix, λ is the regularization coefficient, and I is the identity matrix; or a constrained linear / quadratic programming approach can be used: Tmin ≤ T(x) ≤ Tmax, |u_i| ≤ utol, where T is the cable force and u_i is the control point displacement. The solved Δx is the stress-free cable length adjustment for each cable.

[0035] Output of on-site executable quantities: The material cutting length correction is estimated using the first-order approximation ΔS0 ≈ -ΔT·S0 / (EA), and the refined solution is obtained from the inverse kinematics of the catenary; the jack tensioning stroke is calculated by combining the adjustment amount with the geometric stiffness of the cable. Example: The inverse kinematics of the target alignment yields ΔS0 = -12.3mm for cable No. 12, corresponding to a material cutting length correction of -12.3mm and a tensioning stroke of +58mm; the output table provides line by line for cable number, adjustment amount, material cutting correction, tensioning stroke, and verification status. The system automatically verifies the adjustment results using the closure criterion of "one-time bridge completion ≡ phased construction" and "forward assembly ≡ backward dismantling," and only after passing the verification can the approval process begin.

[0036] Alternative implementation methods The number of sub-models can be customized: the minimum combination is a physical sub-model and a symbolic sub-model; the learning sub-model can be a regression, neural network, Gaussian process, or radial basis function surrogate; the physical sub-model solver can be replaced with a self-developed computational kernel, OpenSees, or a commercial solver, and the method itself does not depend on a specific solver; creep handling can use age-adjusted effective modulus (AAEM) one-step equivalent modulus, or Kelvin chain exponential recursion; the verification granularity can be coarse or fine, and the tolerance can be configured; the conclusion object can be expanded to multi-level verification (single / double person).

[0037] The deployment mode can be local or cloud-based: in the cloud mode, the engineering simulation model and orchestration executor are deployed on a remote server, and the client calls a unified interface via the network; the application domain is switched by configuration: the newly created design domain adopts the design specifications and deliverable templates, and the old bridge reinforcement domain adopts the reinforcement specifications and evaluation deliverable templates, sharing the same engineering simulation model mechanism.

[0038] Electronic devices and media: Electronic devices include a processor and a memory, the memory storing a computer program, and the processor executing the program to implement any of the above steps; a computer program stored on a computer-readable storage medium, when executed by a processor, implements any of the above steps. The above embodiments can be combined arbitrarily, and any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A bridge intelligent analysis method based on a bridge engineering simulation model, characterized in that: Includes the following steps: Step 1: Construct a bridge engineering simulation model: register a physical sub-model, a learning sub-model, a symbolic sub-model, and a state evolution sub-model; wherein, the physical sub-model provides accurate response prediction based on a deterministic numerical calculation kernel; the learning sub-model provides approximate prediction with uncertainty based on a surrogate model; the symbolic sub-model provides constraints and their sources based on a knowledge base of standard provisions; and the state evolution sub-model describes the state transitions of the bridge throughout its entire life cycle based on the construction stage sequence. Step 2: Provide a unified interface: encapsulate five types of operations: prediction, parameter change deduction, scheme deduction, conclusion verification, and actual measurement calibration. This ensures that the predictions, deductions, and conclusions of the automated processing unit and the user interface are all executed or verified through this unified interface, avoiding bypass channels that circumvent physical and regulatory constraints. Step 3, Ordered Execution: The automated processing unit selects the workflow through the function registry, generates the execution sequence by sorting the dependency graph topology, and executes it step by step by the ordered executor. Before each step is executed, a precondition check is performed, and after each step is executed, a postcondition check is performed. In the exploration phase, the learning sub-model is forced to be used and uncertainties are labeled with surrogate estimation. In the issuance phase, the physical sub-model is forced to be used for precise verification, and the direct generation of conclusion objects from approximate results is prohibited. Step 4, Reproducibility Verification: Verify whether the values ​​output by the automated processing unit can be reproduced in the bridge engineering simulation model; the verification is based on the location-working condition-unit-component quadruple, checking the running identifier and model hash corresponding to the values. If they cannot be reproduced, they are blocked and not included in the conclusion object. Step 5, Issuance and Calibration: Generate a conclusion object containing the conclusion, supporting clauses, margin, operation identifier, model hash, and manual review status; the automated processing unit can only produce the status pending review, which takes effect after manual review; and the bridge engineering simulation model parameters are corrected in reverse using measured monitoring data and measured calibration operations, and then proceed to the next round of prediction.

2. The intelligent bridge analysis method based on a bridge engineering simulation model according to claim 1, characterized in that: In step one, the rules of the symbolic sub-model carry the specification version field and clause location, and automatically identify expired rules when the specification version changes.

3. The intelligent bridge analysis method based on a bridge engineering simulation model according to claim 2, characterized in that: In step one, the state evolution sub-model uses the stress-free length and stress-free curvature of the unit as physical anchor points: the model is independently reconstructed and solved for each construction stage, and the newly installed units are assembled into the reference configuration according to the explicitly given stress-free state quantities; and the consistency between the results of one-time bridge completion and phased construction, and the consistency between the results of forward assembly and backward dismantling are used as closure criteria for automatic verification.

4. The intelligent bridge analysis method based on a bridge engineering simulation model according to claim 3, characterized in that: In step one, an influence matrix is ​​constructed using the stress-free cable length as a design variable. The stress-free cable length adjustment corresponding to the target bridge completion state is then solved, and the material cutting length correction and jack tensioning stroke are output.

5. The intelligent bridge analysis method based on a bridge engineering simulation model according to claim 4, characterized in that: In step two, the parameter change deduction is used to return the difference vector of the parameter change relative to the baseline response and its associated components; the scheme deduction is used to deduce the candidate action sequence into a state trajectory that evolves with the construction stage.

6. The intelligent bridge analysis method based on a bridge engineering simulation model according to claim 5, characterized in that: The entries in the function registry in step three include input / output contracts, preconditions, precision levels, and applicable bridge types; adding a new function means registering a line and making the automated processing unit automatically visible; the language model is only responsible for workflow selection and parameter suggestions, and has no right to change the execution order.

7. The intelligent bridge analysis method based on a bridge engineering simulation model according to claim 6, characterized in that: The conclusion object in step three includes the conclusion, the supporting clauses, the margin, the running identifier, the model hash, and the manual review status; the transition from the pending review status to the effective status requires manual action.

8. The intelligent bridge analysis method based on a bridge engineering simulation model according to claim 7, characterized in that: The bridge application domain includes the new bridge design domain and the old bridge reinforcement domain, with different specifications, verification procedures and deliverable templates configured to distinguish the different application domains.

9. A system employing the intelligent bridge analysis method based on a bridge engineering simulation model as described in any one of claims 1 to 8, characterized in that: It includes a processor and a memory, the memory storing a computer program, and the processor executing the computer program to implement the above-described intelligent bridge analysis method.

Images