Method for evaluating bridge operation state by deep learning surrogate model based on finite element guidance

Abstract

A method for evaluating a bridge operation state by a deep learning surrogate model based on finite element guidance, which relates to the technical field of bridge structure online simulation. The method comprises: S1, establishing a finite element simulation model; S2, on the basis of the position, license plate information, axle weight, and license plate information of a vehicle, obtaining vehicle load information, on the basis of the temperature, humidity, and experienced wind speed and wind direction information of a bridge, obtaining environment load information, and, on the basis of the vehicle load information and the environment load information, obtaining vehicle and environment load information; S3, on the basis of finite element guidance, performing adaptive training of a finite element surrogate model; and S4, inputting the vehicle and environment load information into a deep learning neural network surrogate model based on finite element guidance, and outputting a real-time structural state of bridge operation.

Description

A method for evaluating bridge operation status based on finite element guided deep learning surrogate model Technical Field

[0001] The present application relates to a method for evaluating the operating state of a bridge structure, and in particular to a method for evaluating the operating state of a bridge based on a finite element guided deep learning agent model, and belongs to the technical field of online simulation of bridge structures. Background Art

[0002] The number of bridges in existence is increasing, and traffic volume is also increasing. The operation and maintenance supervision of bridges is becoming increasingly important. How to ensure the safe operation of bridges is an area that needs to be studied and explored in the future. In fact, it is not just bridges that face this problem. Other transportation facilities, such as tunnels and slopes, also face this problem. The safety monitoring of bridge facilities during operation is a very important field. This field has also made certain progress in recent years, but the current technology is still at the level of alarming and trend analysis of monitoring thresholds. In addition, the simulation model of the bridge during operation is only established during the monitoring plan design phase and can only simulate a single current situation. How to achieve real-time online simulation of bridge structures is a key research direction.

[0003] Some researchers have proposed a method and device for optimizing the finite element iterative process based on deep learning, CN114021414B; this method optimizes the computational efficiency and convergence of the finite element model on the basis of using deep learning algorithms to predict the next state point in the iteration, reducing the number of iteration steps, and ensuring the accuracy of the simulation results. It has high versatility, but this method is based on a data-driven deep learning network, which requires manual preparation of a large amount of data to train the model, which takes a lot of time; without understanding the data input and data relationship, the reliability of the trained model is poor, and there is no evaluation of the operating structural status of the bridge during the operation period.

[0004] Application Contents

[0005] A brief overview of the present application is provided below to provide a basic understanding of certain aspects of the present application. It should be understood that this overview is not an exhaustive overview of the present application. It is not intended to identify key or important portions of the present application, nor is it intended to limit the scope of the present application. Its purpose is simply to present certain concepts in a simplified form as a prelude to the more detailed description that will be discussed later.

[0006] In view of this, and to address the technical problems existing in the prior art, this application provides a method for evaluating the operating status of a bridge based on a finite element-guided deep learning agent model. Compared with purely data-driven neural network learning, this application imposes physical information constraints during the training process, thereby learning a more generalizable model with fewer data samples. Compared with traditional deep learning, the finite element-guided neural network can autonomously complete model training through physical guidance. It can solve the problems of time-consuming and low simulation efficiency of finite element model simulation, realize real-time simulation of bridges, and evaluate the operating structural status of bridges.

[0007] Solution 1: A method for evaluating bridge operating conditions based on a finite element-guided deep learning agent model, including the following steps:

[0008] S1. Establish a finite element simulation model;

[0009] S2. Vehicle load information is obtained based on the vehicle position, license plate information, axle weight, and license plate information. Environmental load information is obtained based on the bridge's temperature, humidity, and wind speed and direction information. Vehicle environmental load information is obtained based on the vehicle load information and environmental load information.

[0010] S3. Adaptive training of finite element surrogate models based on finite element guidance;

[0011] S4. Input the vehicle environmental load information into the deep learning neural network agent model based on finite element guidance. The model calculates the data and outputs the real-time structural status of the bridge operation.

[0012] In some embodiments, the method of establishing a finite element simulation model is:

[0013] S11. Obtaining bridge structure dimensions and material type parameters from the bridge design drawing;

[0014] S12. Use the finite element method to establish a bridge finite element model that includes structural geometry and material information;

[0015] Among them, u is the structural response displacement parameter, E is the structural elastic modulus, I is the section polar moment of inertia parameter, B is the second derivative of the shape function, N is the shape function, x is the structural longitudinal position parameter, and L is the longitudinal length of the bridge structure.

[0016] In some embodiments, vehicle load information is obtained based on vehicle position, license plate information, axle weight, and license plate information, and environmental load information is obtained based on bridge temperature, humidity, and wind speed and direction information. A method for obtaining vehicle environmental load information based on vehicle load information and environmental load information is:

[0017] S21. Obtain vehicle location and license plate information;

[0018] S22. Acquisition of axle weight and license plate information;

[0019] S23. Obtain the license plate, axle weight, location information and time information of each vehicle by marking the information, and jointly construct the axle weight position distribution information of vehicles passing through the bridge over time as vehicle load information;

[0020] S24. Using the front-end temperature, humidity, wind speed and direction sensors to collect real-time environmental load information of the bridge, including the bridge's temperature, humidity, and wind speed and direction information, to jointly construct the time-varying environmental load information of the bridge as environmental load information;

[0021] S25. Match the vehicle load information and the environmental load information using the time information contained therein to obtain the vehicle environmental load information.

[0022] In some embodiments, the method for obtaining vehicle location and license plate information includes the following steps:

[0023] S211. Deploy video capture equipment on the bridge to capture video information from all lanes on the bridge deck;

[0024] S212. The video capture device captures each frame of video and uses a deep learning target recognition algorithm to identify the license plate and vehicle model information of vehicles on the bridge in real time. The license plate information is used as the vehicle annotation information.

[0025] S213. Within the acquisition area, the longitudinal position and the lateral position are divided according to the lane direction and the lane normal direction, and are represented by x and y, respectively, to construct a lane coordinate system.

[0026] S214. The horizontal and vertical positions in the video image are represented by Ox and Oy to construct an image coordinate system;

[0027] S215. The lane coordinate system and the image coordinate system are transformed using the spatial coordinate transformation equation based on the positional relationship between the camera and the lane:

[0028] x=Ox·cosθ-Oy·sinθ

[0029] y=Ox·sinθ+Oy·cosθ

[0030] S216. For vehicles that appear in the video acquisition area, the position of the vehicle in the image is determined based on the position of the center of the vehicle head, and the position of the vehicle in the lane coordinate system, that is, the actual position information, is obtained. At the same time, the time information of each frame of the image captured by the camera is obtained. The position and time information of each vehicle are distinguished by the license plate.

[0031] In some embodiments, the method for obtaining axle weight and license plate information is: laying a lane-level dynamic weighing system on each lane at the front end of the bridge; using the camera and weighing equipment in the dynamic weighing system to collect the vehicle's license plate and axle weight information, the license plate is used as the marking information, and the axle weight information of each vehicle is distinguished by the license plate.

[0032] In some embodiments, a method for adaptive training of a finite element proxy model based on finite element guidance includes the following steps:

[0033] S31. Initialize model parameters: Define the network structure of the deep learning neural network based on the number of model input variables, the scale of the bridge finite element model, and the number of output variables required, including the number of input and output neurons, the number of neural network layers, and the number of neurons in each layer;

[0034] S32. Dataset construction: The vehicle environmental load information data collected for one day is input into the model as data set a; the vehicle environmental load information data collected for ten days is input into the model as data set b; the data structure of the vehicle environmental load information data collected for three days is disrupted and input into the model as a test set a;

[0035] S33. Define the loss function:

[0036] Add the loss function of the finite element domain:

[0037] L total =w data L data +w Energy L Energy

[0038] Data-based loss function:

[0039] The potential-based loss function is as follows:

[0040] S34. Define an optimization algorithm using the quasi-Newton method in the second-order optimization method;

[0041] S35. Model iterative training: Input dataset a and dataset b into L data and L Energy Adaptive iterative training is performed in the two-part loss function until the loss value reaches a value lower than the allowable value, which is set to 0.00001. After reaching the allowable value, training is automatically stopped and a deep learning neural network proxy model based on finite element guidance is output;

[0042] S36. Evaluate the accuracy of the finite element-guided deep learning neural network proxy model trained using the test set a data input. The evaluation method is to manually check some data. When the checked data reaches 99.5% of the simulation accuracy, the model training is successful. Otherwise, return to S35 to re-iterate the model, and the allowable value will be reduced by 10% until the accuracy requirements are met.

[0043] In some embodiments, a method for inputting vehicle environmental load information into a finite element-guided deep learning neural network agent model, which calculates the data and outputs the real-time structural status of a bridge, includes the following steps:

[0044] S41. Input the vehicle environmental load information into a finite element-guided deep learning neural network agent model for calculation. This involves using trained neurons to perform superposition calculations and calculate output parameters. The output parameters are response parameters of the bridge structure, including deflection, stress, and strain parameters at key locations on the bridge.

[0045] S42. Obtain the maximum deflection ω of the bridge structure response parameter;

[0046] The calculation is carried out according to the provisions of the bridge design code on the deflection limit value, as follows:

[0047] Among them, ω limit Indicates the disturbance limit value, η θ represents the long-term growth coefficient of deflection, and L represents the structural length of the bridge;

[0048] The bridge condition is evaluated based on the ratio of the maximum deflection of the bridge to the deflection limit:

[0049] Among them, η w Represents the bridge state value, when η w When η is between 0.5 and 0.8, the bridge is in a safe state. w When it is between 0.8 and 1.0, the bridge is in a dangerous state. w When it is greater than 1, the bridge status is dangerous.

[0050] Solution 2: An electronic device includes a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, it implements the steps of the method for evaluating the operating status of a bridge based on a finite element-guided deep learning agent model as described in Solution 1.

[0051] Option 3: A computer-readable storage medium storing a computer program, which, when executed by a processor, implements the method for evaluating the operating status of a bridge based on a finite element-guided deep learning agent model as described in Option 1.

[0052] The beneficial effects of this application are as follows: The deep learning agent model based on finite element guidance established in this application realizes real-time online simulation of bridge status under actual environment and vehicle load. The simulation speed can reach millisecond level, truly realizing real-time diagnosis of bridge status. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0054] Figure 1 is a flow chart of the method for evaluating bridge operating status based on finite element guided deep learning agent model. DETAILED DESCRIPTION

[0055] In order to make the technical solutions and advantages of the embodiments of the present application more clearly understood, the exemplary embodiments of the present application are further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, and are not an exhaustive list of all the embodiments. It should be noted that the embodiments and features in the embodiments of the present application can be combined with each other unless they conflict.

[0056] Example 1, referring to FIG1 , illustrates this embodiment, a real-time online bridge simulation method, comprising the following steps:

[0057] S1. Establish finite element simulation module;

[0058] S11. Obtaining bridge structure dimensions and material type parameters from the bridge design drawing;

[0059] S12. Use the finite element method to establish a bridge finite element model that includes structural geometry and material information;

[0060] Among them, u is the structural response displacement parameter, E is the structural elastic modulus, I is the section polar moment of inertia parameter, B is the second derivative of the shape function, N is the shape function, x is the structural longitudinal position parameter, and L is the longitudinal length of the bridge structure.

[0061] S2. Vehicle load information is obtained based on the vehicle position, license plate information, axle weight, and license plate information. Environmental load information is obtained based on the bridge's temperature, humidity, and wind speed and direction information. Vehicle environmental load information is obtained based on the vehicle load information and environmental load information.

[0062] S21. Obtain vehicle location and license plate information;

[0063] S211. Deploy video capture equipment on the bridge to capture video information from all lanes on the bridge deck;

[0064] S212. The video capture device captures each frame of video and uses a deep learning target recognition algorithm to identify the license plate and vehicle model information of vehicles on the bridge in real time. The license plate information is used as the vehicle annotation information.

[0065] S213. Each video capture device captures a fixed area. Within the capture area, the longitudinal position and the lateral position are divided according to the lane direction and lane normal direction, respectively represented by x and y, to construct a lane coordinate system.

[0066] S214. Since the camera position does not necessarily match the horizontal and vertical positions of the lane, the horizontal and vertical positions in the video image are represented by Ox and Oy to construct an image coordinate system.

[0067] S215. The lane coordinate system and the image coordinate system are transformed using the spatial coordinate transformation equation based on the positional relationship between the camera and the lane:

[0068] x=Ox·cosθ-Oy·sinθ

[0069] y=Ox·sinθ+Oy·cosθ

[0070] S216. For vehicles within the video capture area, the position of the vehicle in the image is determined based on the center of the vehicle's head. This determines the vehicle's position in the lane coordinate system, i.e., its actual position information. The time information of each frame captured by the camera is also obtained. The position and time information of each vehicle are distinguished by the license plate.

[0071] S22. Acquisition of axle weight and license plate information;

[0072] S221. Install a lane-level dynamic weighing system on each lane at the front end of the bridge. The system uses cameras and weighing equipment in the dynamic weighing system to collect vehicle license plate and axle weight information, with the license plate as the annotation information. Each vehicle's axle weight information is distinguished by the license plate.

[0073] S23. Obtain the license plate, axle weight, location information and time information of each vehicle by marking the information, and jointly construct the axle weight position distribution information of vehicles passing through the bridge over time as vehicle load information;

[0074] S24. Using the front-end temperature, humidity, wind speed and direction sensors to collect real-time environmental load information of the bridge, including the bridge's temperature, humidity, and wind speed and direction information, to jointly construct the time-varying environmental load information of the bridge as environmental load information;

[0075] S25 matches the vehicle load information and the environmental load information using the time information contained therein to obtain the vehicle environmental load information.

[0076] S3. Adaptive training of finite element surrogate models based on finite element guidance;

[0077] S31. Initialize model parameters: Define the network structure of the deep learning neural network based on the number of model input variables, the scale of the bridge finite element model, and the number of output variables required, including the number of input and output neurons, the number of neural network layers, and the number of neurons in each layer;

[0078] S32. Dataset construction: Input the vehicle environmental load information data collected for one day as dataset a into the model; input the vehicle environmental load information data collected for ten days as dataset b into the model; and scramble the data structure of the vehicle environmental load information data collected for three days and input it into the model as test set a.

[0079] S33. Define the loss function:

[0080] Add the loss function of the finite element domain:

[0081] L total =w data L data +w Energy L Energy

[0082] Data-based loss function:

[0083] The potential-based loss function is as follows:

[0084] S34. Define an optimization algorithm using the quasi-Newton method in the second-order optimization method;

[0085] S35. Model iterative training: Input dataset a and dataset b into L data and L Energy Adaptive iterative training is performed in the two-part loss function until the loss value reaches a value lower than the allowable value, which is set to 0.00001. After reaching the allowable value, training is automatically stopped and a deep learning neural network proxy model based on finite element guidance is output;

[0086] S36. Evaluate the accuracy of the finite element-guided deep learning neural network proxy model trained using the test set a data input. The evaluation method is to manually check some data. When the checked data reaches 99.5% of the simulation accuracy, the model training is successful. Otherwise, return to S35 to re-iterate the model, and the allowable value will be reduced by 10% until the accuracy requirements are met.

[0087] S4. Input the vehicle environmental load information into the deep learning neural network agent model based on finite element guidance. The model calculates the data and outputs the real-time structural status of the bridge operation.

[0088] S41. Input the vehicle environmental load information into a finite element-guided deep learning neural network proxy model for calculation. This involves utilizing trained neurons to perform superposition calculations and calculate output parameters. The output parameters are response parameters of the bridge structure, including deflection, stress, and strain parameters at key locations on the bridge.

[0089] S42. Obtain the maximum deflection ω of the bridge structure response parameter;

[0090] The calculation is carried out according to the provisions of the bridge design code on the deflection limit value, as follows:

[0091] Among them, ω limit Indicates the disturbance limit value, η θ represents the long-term growth coefficient of deflection, and L represents the structural length of the bridge;

[0092] The bridge condition is evaluated based on the ratio of the maximum deflection of the bridge to the deflection limit:

[0093] Among them, η w Represents the bridge state value, when η w When η is between 0.5 and 0.8, the bridge is in a safe state. w When it is between 0.8 and 1.0, the bridge is in a dangerous state. w When the value is greater than 1, the bridge is in a dangerous state. Example 2: The computer device of the present application may include a processor and memory, such as a single-chip microcomputer including a central processing unit. Furthermore, the processor is configured to execute a computer program stored in the memory to implement the steps of the aforementioned method for assessing bridge operating conditions based on a finite element-guided deep learning agent model.

[0094] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0095] The memory may mainly include a program storage area and a data storage area. The program storage area may store an operating system and at least one application required for a function (such as a sound playback function, an image playback function, etc.); the data storage area may store data created based on the use of the mobile phone (such as audio data, a phone book, etc.). In addition, the memory may include a high-speed random access memory and may also include a non-volatile memory, such as a hard disk, internal memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state storage device.

[0096] Embodiment 3, computer readable storage medium embodiment.

[0097] The computer-readable storage medium of the present application can be any form of storage medium that can be read by the processor of a computer device, including but not limited to non-volatile memory, volatile memory, ferroelectric memory, etc. The computer-readable storage medium stores a computer program. When the processor of the computer device reads and executes the computer program stored in the memory, the steps of the above-mentioned method for evaluating the operating status of a bridge based on a finite element-guided deep learning agent model can be implemented.

[0098] The computer program includes computer program code, which may be in source code form, object code form, executable file, or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal, and software distribution medium. It should be noted that the content contained in the computer-readable medium may be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0099] Although the present application has been described in terms of a limited number of embodiments, it will be apparent to those skilled in the art, having benefit of the foregoing description, that other embodiments are contemplated within the scope of the present application as described herein. Furthermore, it should be noted that the language used in this specification has been selected primarily for readability and didactic purposes, rather than for the purpose of explaining or limiting the subject matter of the present application. Consequently, many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the appended claims. The disclosure made herein is illustrative and non-restrictive of the scope of the present application, and the scope of the present application is defined by the appended claims.

Claims

1. A method for evaluating bridge operation status based on finite element guided deep learning agent model, characterized in that: The following steps are involved: S1. Establish finite element simulation model; S2. Obtain vehicle load information based on vehicle position, license plate information, axle weight, and license plate information, obtain environmental load information based on bridge temperature, humidity, and wind speed and direction information, and obtain vehicle environmental load information based on vehicle load information and environmental load information; S3. Adaptive training of finite element surrogate models based on finite element guidance; S4. Input the vehicle environmental load information into the deep learning neural network agent model guided by finite elements. The model calculates the data and outputs the real-time structural status of the bridge operation.

2. According to claim 1, the method for evaluating bridge operation status based on finite element guided deep learning agent model is characterized in that: The method to establish the finite element simulation model is: S11. Obtaining bridge structure dimensions and material type parameters on the bridge design drawing; S12. Establish a finite element model of the bridge including structural geometry and material information by using the finite element method; Among them, u is the structural response displacement parameter, E is the structural elastic modulus, I is the section polar moment of inertia parameter, B is the second derivative of the shape function, N is the shape function, x is the structural longitudinal position parameter, and L is the longitudinal length of the bridge structure.

3. The method for evaluating bridge operation status based on finite element guided deep learning agent model according to claim 2 is characterized in that: The vehicle load information is obtained according to the vehicle position, license plate information, axle weight and license plate information, and the environmental load information is obtained according to the bridge temperature, humidity and wind speed and direction information. The method for obtaining the vehicle environmental load information based on the vehicle load information and environmental load information is: S21. Obtain vehicle location and license plate information; S22. Acquisition of axle weight and license plate information; S23. Obtain the license plate, axle weight, location information and time information of each vehicle through the labeling information, and jointly construct the axle weight position distribution information of the vehicles passing through the bridge deck over time as the vehicle load information; S24. Using the temperature and humidity, wind speed and direction sensors at the front end to collect the environmental load information of the bridge in real time, including the temperature, humidity and wind speed and direction information of the bridge, and jointly construct the environmental load information of the bridge that changes over time as the environmental load information; S25. Match the vehicle load information and the environmental load information by using the time information contained in both. As vehicle environmental load information.

4. The method for evaluating bridge operation status based on finite element guided deep learning agent model according to claim 3 is characterized in that: The method for obtaining vehicle position and license plate information comprises the following steps: S211. Deploy video acquisition equipment on the bridge to collect video information of all lanes on the bridge deck; S212. The video acquisition device collects each frame of video images and uses the target recognition deep learning algorithm to identify the license plate and model information of the vehicle on the bridge in real time, and the license plate information is used as the vehicle annotation information; S213. In the acquisition area, the longitudinal position and the lateral position are divided according to the lane direction and the lane normal direction, and are represented by x and y respectively, to construct a lane coordinate system; S214. The horizontal position and the vertical position in the video image are represented by Ox and Oy to construct an image coordinate system; S215. The lane coordinate system and the image coordinate system are transformed by using the spatial coordinate transformation equation based on the positional relationship between the camera and the lane: x=Ox·cosθ-Oy·sinθ y+Ox·sinθ+Oy·cosθ S216. For vehicles that appear in the video acquisition area, the position of the vehicle in the image is determined based on the position of the center of the vehicle head, and the position of the vehicle in the lane coordinate system, that is, the actual position information, is obtained. At the same time, the time information of each frame of the image captured by the camera is obtained. The position and time information of each vehicle are distinguished by the license plate.

5. The method for evaluating bridge operation status based on finite element guided deep learning agent model according to claim 4 is characterized in that: The method for obtaining axle weight and license plate information is: laying a lane-level dynamic weighing system on each lane at the front end of the bridge; using the camera and weighing equipment in the dynamic weighing system to collect the vehicle's license plate and axle weight information, with the license plate as the marking information, and the axle weight information of each vehicle is distinguished by the license plate.

6. The method for evaluating bridge operation status based on finite element guided deep learning agent model according to claim 5 is characterized in that: The method for adaptive training of a finite element proxy model based on finite element guidance comprises the following steps: S31. Initialize model parameters: define the network structure of the deep learning neural network according to the number of model input variables, the scale of the bridge finite element model and the number of output variables, including the number of input and output neurons, the number of neural network layers and the number of neurons in each layer; A32. Dataset construction: The vehicle environment load information data collected for one day is input into the model as data set a; the vehicle environment load information data collected for ten days is input into the model as data set b; the data structure of the vehicle environment load information data collected for three days is disrupted and input into the model as test set a; S33. Define the loss function: Add the loss function of the finite element domain: L total =w data L data +w Energy L Energy Data-based loss function: The potential-based loss function is as follows: S34. Define an optimization algorithm using the quasi-Newton method in the second-order optimization method; S35. Model iterative training: Input dataset a and dataset b into L data and L Energy Adaptive iterative training is performed in the two-part loss function to make the loss value reach a value lower than the allowable value. The allowable value is set to 0.00001. After reaching the allowable value, the training is automatically stopped and the deep learning neural network proxy model guided by finite elements is output; S36. The finite element-guided deep learning neural network proxy model trained by inputting the test set a data is evaluated for its accuracy. The evaluation method is to manually check some data. When the checked data reaches 99.5% of the simulation accuracy, the model training is successful. Otherwise, return to S35 to re-iterate the model, and the allowable value will be reduced by 10% until the accuracy requirement is met.

7. The method for evaluating bridge operation status based on finite element guided deep learning agent model according to claim 6 is characterized in that: The vehicle environmental load information is input into a deep learning neural network agent model based on finite element guidance, and the model calculates the data and outputs a method for obtaining the real-time structural state of the bridge operation, including the following steps: S41. Input the vehicle environmental load information into the deep learning neural network agent model based on finite element guidance for calculation, that is, use the trained neurons to perform superposition calculation and calculate the output parameters, which are the response parameters of the bridge structure. The response parameters include the deflection, stress and strain parameters of the key positions of the bridge; S42. Obtain the maximum deflection ω of the bridge structure response parameter; The calculation is carried out according to the provisions of the deflection limit value in the bridge design specifications, as follows: Among them, ω limit Indicates the disturbance limit value, η θ represents the long-term growth coefficient of deflection, and L represents the structural length of the bridge; The bridge condition is evaluated based on the ratio of the maximum deflection of the bridge to the deflection limit value: Among them, η w Represents the bridge state value, when η w When η is between 0.5 and 0.8, the bridge is in a safe state. w When it is between 0.8 and 1.0, the bridge is in a dangerous state. w When it is greater than 1, the bridge is in a dangerous state.

8. An electronic device, characterized in that: It includes a memory and a processor, the memory stores a computer program, and the processor implements the steps of the method for evaluating the operating status of a bridge based on a finite element guided deep learning agent model as described in any one of claims 1 to 7 when executing the computer program.

9. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed by a processor, the method for evaluating the operating status of a bridge based on a finite element guided deep learning agent model as described in any one of claims 1 to 7 is implemented.

Images