Crack evaluation method and system based on mass concrete bundle stress analysis

Abstract

The invention discloses a crack evaluation method and system based on mass concrete beam stress analysis, and relates to the technical field of concrete crack evaluation.The method comprises the steps that sample data are input into an initial model based on deep learning in groups to be processed, and on the basis of output of each neuron of each layer of network in the initial model, the initial model is constructed; updating the preset adjustable parameters in the initial model and the connection strength between the neurons; calculating a loss function of each group of input data according to the output of the initial model; determining the initial model as a crack evaluation model until the loss function reaches a preset training end condition; acquiring parameter data of a to-be-assessed site, and inputting the parameter data of the to-be-assessed site into the crack assessment model for processing to obtain a crack assessment result of the to-be-assessed site; compared with a traditional mode, the mode is higher in efficiency and accuracy, does not depend on artificial experience, and achieves the purpose of reducing the crack risk of the mass concrete.

Description

Technical Field

[0001] The present invention relates to the technical field of concrete crack assessment, and more particularly to a crack assessment method and system based on mass concrete beam stress analysis. Background Art

[0002] With the rapid development of my country's modern economy and society, demands for multifunctional and integrated urban development are growing in all regions, and the exterior dimensions of buildings are also increasing. Consequently, large-volume concrete structures are increasingly being used in large commercial and public buildings. Cracks are a common problem in concrete structures. Severe cracks can deteriorate the strength and stability of concrete structures, undermining their integrity and impermeability, significantly reducing their bearing capacity and jeopardizing their safe operation. Minor cracks can affect the durability and aesthetics of buildings, and some can even develop into hazardous cracks.

[0003] Concrete cracking and damage has always been a difficult problem to overcome and control in the engineering community. Due to the effects of concrete hydration heat release, shrinkage deformation, loads, environmental erosion, and steel corrosion, cracking is a common phenomenon in reinforced concrete structures. Concrete cracks have a serious impact on the safety and durability of reinforced concrete structures. Therefore, regular monitoring of concrete structures is necessary to implement preventive measures. Currently, most methods rely on manual calculation of relevant parameters and empirical evaluation. This method lacks efficiency and accuracy, and relies on manual experience, resulting in significant limitations in implementation. Summary of the Invention

[0004] The object of the present invention is to provide a crack assessment method and system based on mass concrete beam stress analysis to solve the problems existing in the above-mentioned background technology.

[0005] The above technical objectives of the present invention are achieved through the following technical solutions: In a first aspect, the present application provides a crack assessment method based on mass concrete beam stress analysis, comprising the following specific steps: The sample data is obtained through on-site measurement or simulation using finite element simulation software. The sample data includes multiple sets of parameter data and the corresponding construction results of each set. The parameter data includes time-varying parameters, material parameters, and environmental parameters. The sample data are input into the initial model based on deep learning for processing in groups, and the preset adjustable parameters in the initial model and the connection strength between each neuron are updated based on the output of each neuron in each layer of the network in the initial model; Based on the updated preset adjustable parameters and connection strengths, the loss function for each set of input data is calculated according to the output of the initial model; Until the loss function reaches a preset training end condition, the initial model whose loss function reaches the preset training end condition is determined as the crack assessment model; Obtain parameter data of the site to be assessed, and input the parameter data of the site to be assessed into the crack assessment model for processing to obtain the crack assessment results of the site to be assessed. The crack assessment results include high risk, medium risk and low risk.

[0006] On the basis of the above technical solution, the present invention can also be improved as follows.

[0007] Furthermore, the above loss function is specifically: ,in: ; Where, represents the hyperparameter, ; represents the stress prediction result output by the crack assessment model, represents the true self-restraint stress, represents the true external constraint stress, represents the true stress, Indicates the number of groups of input data.

[0008] Furthermore, the above true stress is calculated by the true self-constraint stress and the true external constraint stress, and the true self-constraint stress and the true external constraint stress are calculated by parameter data, specifically: ,in: , ; Where, represents the stress prediction result output by the crack assessment model, represents the true self-restraint stress, represents the true external constraint stress, represents the time-varying elastic modulus, is the expansion coefficient, is the temperature difference between the inside and outside of the cross section, is the relaxation coefficient, represents the constraint stiffness ratio, represents the temperature shrinkage strain, Indicates drying shrinkage strain.

[0009] Furthermore, the preset adjustable parameters in the initial model and the connection strength between each neuron are updated as follows: After each set of sample data is input into the initial model, the output of each neuron in each layer of the network in the initial model is calculated; Based on the output of each neuron and the preset adjustable parameters, the gradient of the output of each neuron is calculated, and based on the gradient of the output of each neuron and the preset adjustable parameters, the gradient of the connection strength between each neuron is calculated; Based on the gradient of the connection strength between each neuron, the connection strength between each neuron is updated, and after the connection strength between each neuron is updated, the preset adjustable parameters are readjusted.

[0010] Furthermore, the output of each neuron in each layer of the above network is specifically: ; Where, Indicates the t The first iteration l Neurons in a layer network i The output, Indicates the l Neurons in a layer network i and neurons j The connection strength between represents the activation function, and the neuron i After output ， ,otherwise ; represents the parameter that controls the change of neuronal membrane potential over time, Indicates the t This iteration l Neurons in a layer network i The membrane potential before output, Indicates the t This iteration l Neurons in a layer network i Input, represents the upper threshold of membrane potential, Represents the Heaviside function.

[0011] Furthermore, the gradient of the output of each of the above neurons is specifically: ,in: , ; Where, Indicates the l Neurons in a layer network i The gradient of the output of Indicates the l- Neurons in a 1-layer network i The gradient of the output of Indicates the l Neurons in a layer network iThe gradient of the input, represents the loss function of connection strength, , A represents the intermediate quantity along the spatial direction, B represents the intermediate quantity along the time direction, All represent preset adjustable parameters. Indicates the t This iteration l Neurons in a layer network i The membrane potential before output, Indicates the t This iteration l Neurons in a layer network i The membrane potential after output, represents the upper threshold of membrane potential, represents the parameter that controls the temporal changes of neuronal membrane potential; The gradient of the connection strength between neurons is specifically: ; Where, Indicates the l Neurons in a layer network i and neurons j The gradient of the connection strength between Indicates the l Neurons in a layer network i The gradient of the output of All represent preset adjustable parameters. Indicates the t This iteration l Neurons in a layer network i Input, Indicates the t This iteration l Neurons in a layer network i The membrane potential after output, represents the upper threshold of membrane potential; Update the connection strength between neurons, specifically: ; Where, No. l Neurons in a layer network i and neurons j The change in the connection strength between represents the learning rate of the model, Indicates the l Neurons in a layer network i and neurons j The gradient of the connection strength between A loss function representing the strength of the connection.

[0012] Furthermore, the above preset adjustable parameters are specifically: , ; Where, All represent preset adjustable parameters. Indicates preset adjustable parameters The minimum value of Indicates preset adjustable parameters The maximum value of T represents the maximum number of iterations, t Indicates the current iteration number.

[0013] In a second aspect, the present application provides a crack assessment system based on mass concrete beam stress analysis, which is applied to any one of the crack assessment methods based on mass concrete beam stress analysis in the first aspect, comprising: The first module is used to obtain sample data through on-site measurement or simulation using finite element simulation software. The sample data includes multiple sets of parameter data and the corresponding construction results of each set. The parameter data includes time-varying parameters, material parameters, and environmental parameters. The second module is used to input the sample data into the initial model based on deep learning for processing in groups, and based on the output of each neuron in each layer of the network in the initial model, update the preset adjustable parameters in the initial model and the connection strength between each neuron; The third module is used to calculate the loss function of each set of input data according to the output of the initial model based on the updated preset adjustable parameters and connection strength; The fourth module is configured to determine the initial model whose loss function reaches the preset training end condition as the crack assessment model until the loss function reaches the preset training end condition; The fifth module is used to obtain parameter data of the site to be evaluated, and input the parameter data of the site to be evaluated into the crack assessment model for processing to obtain the crack assessment results of the site to be evaluated. The crack assessment results include high risk, medium risk and low risk.

[0014] In a third aspect, the present application provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements any one of the methods in the first aspect when executing the computer program.

[0015] In a fourth aspect, the present application provides a non-transitory computer-readable storage medium, which stores computer instructions, and the computer instructions enable a computer to execute any one of the methods in the first aspect.

[0016] Compared with the prior art, the present invention has at least the following beneficial effects: In the present application, first, sample data for training the initial model is obtained based on the measured data of the construction site and the finite element simulation software. The sample data is divided into multiple groups of parameter data and the corresponding construction results of each group; secondly, the initial model is trained using the parameter data of each group and the corresponding construction results of each group. During the training process, the preset adjustable parameters in the initial model and the connection strength between each neuron are updated according to the output of each neuron in each layer of the network. After the preset adjustable parameters and the connection strength are updated, the loss function is calculated again until the loss function reaches the training end condition. The initial model whose loss function reaches the preset training end condition is determined as the crack assessment model; finally, the parameter data of the site to be assessed is used to obtain the crack assessment result of the site to be assessed; the output of the crack assessment model can be a predicted stress, and the ratio of the predicted stress to the standard stress established by the industry is the risk coefficient. When the risk coefficient is greater than 1, it is considered high risk, when the risk coefficient is in the range of 0.8-1, it is medium risk, and when it is less than 0.8, it is considered low risk; compared with the traditional method, this method is more efficient and accurate, and does not rely on manual experience to achieve the purpose of reducing the risk of cracks in large-volume concrete.

[0017] In this application, a neural network model is used for multi-source data fusion, physical law embedding and dynamic analysis, which significantly improves the accuracy and reliability of concrete cracking risk assessment. It is particularly suitable for engineering safety management in complex environments and extreme working conditions. It can maintain or improve the bearing capacity and ductility of concrete structures and structures with existing cracks, and enhance the structure's ability to withstand operation under natural disaster conditions such as earthquakes. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, constitute a part of this application, and do not constitute a limitation of the embodiments of the present invention. In the drawings: Figure 1 A flowchart of an evaluation method according to an embodiment of the present invention; Figure 2 A connection diagram of an evaluation system according to an embodiment of the present invention; Figure 3 Schematic diagram of the connection of electronic equipment in an embodiment of the present invention. DETAILED DESCRIPTION

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations.

[0020] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention as claimed, but rather merely represents selected embodiments of the present invention. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without creative effort shall fall within the scope of protection of the present invention.

[0021] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings.

[0022] Example 1: In order to solve the problem that the current crack assessment of mass concrete relies on manual calculation of relevant parameters and empirical evaluation, which leads to lack of efficiency and accuracy and has great limitations in implementation, this example provides a crack assessment method based on mass concrete beam stress analysis, such as Figure 1 As shown, the following specific steps are included: S1, sample data is obtained through on-site measurement or simulation using finite element simulation software. The sample data includes multiple groups of parameter data and the corresponding construction results of each group. The parameter data includes time-varying parameters, material parameters, and environmental parameters.

[0023] Among them, during on-site measurements, on-site monitoring data can be obtained through sensors, strain gauges, etc., and the construction results of the corresponding data are also required. For example, in the later stage, it is determined through survey whether cracks appear. If cracks appear, they can be classified according to the size of the cracks. Specifically, the above-mentioned time-varying parameters may include elastic modulus, temperature difference, expansion coefficient and relaxation coefficient, etc.; material parameters may include water-cement ratio, fly ash content, etc.; environmental parameters may include ambient temperature and humidity, wind speed, etc., and material parameters and environmental parameters will affect the changes in parameters such as constraint stiffness ratio, temperature shrinkage strain, and drying shrinkage strain.

[0024] S2, inputs the sample data into the initial model based on deep learning for processing in groups, and updates the preset adjustable parameters in the initial model and the connection strength between each neuron based on the output of each neuron in each layer of the network in the initial model.

[0025] Optionally, the preset adjustable parameters in the initial model and the connection strengths between neurons are updated as follows: S21, after each set of sample data is input into the initial model, the output of each neuron in each layer of the network in the initial model is calculated.

[0026] Among them, the output of each neuron in each layer of the above network is specifically: ; Where, Indicates the t The first iteration l Neurons in a layer network i The output, Indicates the l Neurons in a layer network i and neurons j The connection strength between represents the activation function, and the neuron i After output ， ,otherwise ; represents the parameter that controls the change of neuronal membrane potential over time, Indicates the t This iteration l Neurons in a layer network i The membrane potential before output, Indicates the t This iteration l Neurons in a layer network i Input, represents the upper threshold of membrane potential, Represents the Heaviside function.

[0027] S22, based on the output of each neuron and the preset adjustable parameters, calculate the gradient of the output of each neuron, and calculate the gradient of the connection strength between each neuron according to the gradient of the output of each neuron and the preset adjustable parameters.

[0028] Among them, the gradient of the output of each of the above neurons is specifically: ,in: , ; Where, Indicates the l Neurons in a layer network i The gradient of the output of Indicates the l- Neurons in a 1-layer network i The gradient of the output of Indicates the l Neurons in a layer network iThe gradient of the input, represents the loss function of connection strength, , A represents the intermediate quantity along the spatial direction, B represents the intermediate quantity along the time direction, All represent preset adjustable parameters. Indicates the t This iteration l Neurons in a layer network i The membrane potential before output, Indicates the t This iteration l Neurons in a layer network i The membrane potential after output, represents the upper threshold of membrane potential, Represents the parameter that controls the temporal variation of neuronal membrane potential.

[0029] Furthermore, the gradient of the connection strength between the above neurons is specifically: ; Where, Indicates the l Neurons in a layer network i and neurons j The gradient of the connection strength between Indicates the l Neurons in a layer network i The gradient of the output of All represent preset adjustable parameters. Indicates the t This iteration l Neurons in a layer network i Input, Indicates the t This iteration l Neurons in a layer network i The membrane potential after output, Represents the upper threshold of membrane potential.

[0030] S23, based on the gradient of the connection strength between each neuron, the connection strength between each neuron is updated, and after the connection strength between each neuron is updated, the preset adjustable parameters are readjusted; the above-mentioned training end condition can be to set the maximum number of iterations in advance, and each iteration is executed through the above-mentioned steps S21-S23.

[0031] Among them, the connection strength between each neuron is updated, specifically: ; Where, No.l Neurons in a layer network i and neurons j The change in the connection strength between represents the learning rate of the model, Indicates the l Neurons in a layer network i and neurons j The gradient of the connection strength between A loss function representing the strength of the connection.

[0032] Optionally, the above preset adjustable parameters are specifically: , ; Where, All represent preset adjustable parameters. Indicates preset adjustable parameters The minimum value can be 1. Indicates preset adjustable parameters The maximum value can be 10. T represents the maximum number of iterations, t Indicates the current iteration number.

[0033] Among them, the neural network in the initial model has a multi-layer structure. Each iterative process is to adjust the various model parameters of the model, including connection strength, after the change value is input into the initial model. This can ensure that more neurons in the feature learning device participate in learning, so as to achieve higher accuracy after learning.

[0034] S3, based on the updated preset adjustable parameters and connection strength, calculates the loss function for each set of input data according to the output of the initial model.

[0035] The input of data, the update of connection strength, and the calculation of loss function are all part of the model training process. The model is then adjusted in reverse based on the loss function to enable the model to have accurate crack assessment capabilities.

[0036] S4, until the loss function reaches a preset training end condition, the initial model whose loss function reaches the preset training end condition is determined as the crack assessment model.

[0037] Optionally, the above loss function is specifically: ,in: ; Where, represents the hyperparameter, ; represents the stress prediction result output by the crack assessment model, represents the true self-restraint stress, represents the true external constraint stress, represents the true stress, Indicates the number of groups of input data.

[0038] Furthermore, the above true stress is calculated by the true self-constraint stress and the true external constraint stress, and the true self-constraint stress and the true external constraint stress are calculated by parameter data, specifically: ,in: , ; Where, represents the stress prediction result output by the crack assessment model, represents the true self-restraint stress, represents the true external constraint stress, represents the time-varying elastic modulus, is the expansion coefficient, is the temperature difference between the inside and outside of the cross section, is the relaxation coefficient, represents the constraint stiffness ratio, represents the temperature shrinkage strain, Indicates drying shrinkage strain.

[0039] S5, obtaining parameter data of the site to be assessed, and inputting the parameter data of the site to be assessed into a crack assessment model for processing to obtain a crack assessment result of the site to be assessed, wherein the crack assessment result includes high risk, medium risk and low risk.

[0040] Specifically, the crack assessment model can output predicted stress, with the ratio of the predicted stress to the standard stress established by the industry as the risk coefficient. When the risk coefficient is greater than 1, it is considered high risk, when the risk coefficient is in the range of 0.8-1, it is medium risk, and when it is less than 0.8, it is considered low risk. Compared with traditional methods, this method is more efficient and accurate, and does not rely on manual experience, achieving the purpose of reducing the risk of cracks in large-volume concrete.

[0041] Among them, the use of neural network models for multi-source data fusion, physical law embedding and dynamic analysis has significantly improved the accuracy and reliability of concrete cracking risk assessment. It is particularly suitable for engineering safety management in complex environments and extreme working conditions. It can maintain or improve the bearing capacity and ductility of concrete structures and structures with existing cracks, and enhance the structure's ability to withstand operation under natural disaster conditions such as earthquakes.

[0042] Example 2: This embodiment of the present application provides a crack assessment system based on mass concrete beam stress analysis, which is applied to a crack assessment method based on mass concrete beam stress analysis in Example 1, such as Figure 2 Shown, including: The first module is used to obtain sample data through on-site measurement or simulation using finite element simulation software. The sample data includes multiple sets of parameter data and the corresponding construction results of each set. The parameter data includes time-varying parameters, material parameters, and environmental parameters. The second module is used to input the sample data into the initial model based on deep learning for processing in groups, and based on the output of each neuron in each layer of the network in the initial model, update the preset adjustable parameters in the initial model and the connection strength between each neuron; The third module is used to calculate the loss function of each set of input data according to the output of the initial model based on the updated preset adjustable parameters and connection strength; The fourth module is configured to determine the initial model whose loss function reaches the preset training end condition as the crack assessment model until the loss function reaches the preset training end condition; The fifth module is used to obtain parameter data of the site to be evaluated, and input the parameter data of the site to be evaluated into the crack assessment model for processing to obtain the crack assessment results of the site to be evaluated. The crack assessment results include high risk, medium risk and low risk.

[0043] Example 3: This embodiment of the present application provides an electronic device, such as Figure 3 As shown, it includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the method in embodiment 1 is implemented.

[0044] Example 4: An embodiment of the present application provides a non-transitory computer-readable storage medium, which stores computer instructions, and the computer instructions enable a computer to execute the method in Example 1.

[0045] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A crack assessment method based on mass concrete beam stress analysis, characterized in that: The specific steps include: Sample data is obtained by on-site measurement or simulation using finite element simulation software, wherein the sample data includes multiple groups of parameter data and construction results corresponding to each group, wherein the parameter data includes time-varying parameters, material parameters, and environmental parameters; The sample data are input into an initial model based on deep learning for processing in groups, and based on the output of each neuron in each layer of the network in the initial model, the preset adjustable parameters in the initial model and the connection strength between each neuron are updated; Based on the updated preset adjustable parameters and connection strengths, the loss function for each set of input data is calculated according to the output of the initial model; Until the loss function reaches a preset training end condition, the initial model whose loss function reaches the preset training end condition is determined as the crack assessment model; Parameter data of the site to be assessed is obtained, and the parameter data of the site to be assessed is input into a crack assessment model for processing to obtain a crack assessment result of the site to be assessed, wherein the crack assessment result includes high risk, medium risk and low risk.

2. A crack assessment method based on mass concrete beam stress analysis according to claim 1, characterized in that: The loss function is specifically: ,in: ; Where, represents the hyperparameter, ; represents the stress prediction result output by the crack assessment model, represents the true self-restraint stress, represents the true external constraint stress, represents the true stress, Indicates the number of groups of input data.

3. A crack assessment method based on mass concrete beam stress analysis according to claim 2, characterized in that: The true stress is calculated by the true self-constraint stress and the true external constraint stress, and the true self-constraint stress and the true external constraint stress are calculated by parameter data, specifically: ,in: , ; Where, represents the stress prediction result output by the crack assessment model, represents the true self-restraint stress, represents the true external constraint stress, represents the time-varying elastic modulus, is the expansion coefficient, is the temperature difference between the inside and outside of the cross section, is the relaxation coefficient, represents the constraint stiffness ratio, represents the temperature shrinkage strain, Indicates drying shrinkage strain.

4. The crack assessment method based on mass concrete beam stress analysis according to claim 1, characterized in that: Update the preset adjustable parameters in the initial model and the connection strength between each neuron, specifically: After each set of sample data is input into the initial model, the output of each neuron in each layer of the network in the initial model is calculated; Based on the output of each neuron and the preset adjustable parameters, the gradient of the output of each neuron is calculated, and based on the gradient of the output of each neuron and the preset adjustable parameters, the gradient of the connection strength between each neuron is calculated; Based on the gradient of the connection strength between each neuron, the connection strength between each neuron is updated, and after the connection strength between each neuron is updated, the preset adjustable parameter is readjusted.

5. The crack assessment method based on mass concrete beam stress analysis according to claim 4, characterized in that: The output of each neuron in each layer of the network is: ； Where, Indicates the t The first iteration l Neurons in a layer network i The output, Indicates the l Neurons in a layer network i and neurons j The connection strength between represents the activation function, and the neuron i After output ， ,otherwise ; represents the parameter that controls the change of neuronal membrane potential over time, Indicates the t This iteration l Neurons in a layer network i The membrane potential before output, Indicates the t This iteration l Neurons in a layer network i Input, represents the upper threshold of membrane potential, Represents the Heaviside function.

6. The crack assessment method based on mass concrete beam stress analysis according to claim 4, characterized in that: The gradient of the output of each neuron is: ,in: ， ； Where, Indicates the l Neurons in a layer network i The gradient of the output of Indicates the l- Neurons in a 1-layer network i The gradient of the output of Indicates the l Neurons in a layer network i The gradient of the input, represents the loss function of connection strength, , A represents the intermediate quantity along the spatial direction, B represents the intermediate quantity along the time direction, All represent preset adjustable parameters. Indicates the t This iteration l Neurons in a layer network i The membrane potential before output, Indicates the t This iteration l Neurons in a layer network i The membrane potential after output, represents the upper threshold of membrane potential, represents the parameter that controls the temporal changes of neuronal membrane potential; The gradient of the connection strength between neurons is specifically: ； Where, Indicates the l Neurons in a layer network i and neurons j The gradient of the connection strength between Indicates the l Neurons in a layer network i The gradient of the output of All represent preset adjustable parameters. Indicates the t This iteration l Neurons in a layer network i Input, Indicates the t This iteration l Neurons in a layer network i The membrane potential after output, represents the upper threshold of membrane potential; Update the connection strength between neurons, specifically: ； Where, No. l Neurons in a layer network i and neurons j The change in the connection strength between represents the learning rate of the model, Indicates the l Neurons in a layer network i and neurons j The gradient of the connection strength between A loss function representing the strength of the connection.

7. The crack assessment method based on mass concrete beam stress analysis according to claim 4, characterized in that: The preset adjustable parameters are specifically: ， ； Where, All represent preset adjustable parameters. Indicates preset adjustable parameters The minimum value of Indicates preset adjustable parameters The maximum value of T represents the maximum number of iterations, t Indicates the current iteration number.

8. A crack assessment system based on mass concrete beam stress analysis, applied to a crack assessment method based on mass concrete beam stress analysis according to any one of claims 1 to 7, characterized in that: include: The first module is used to obtain sample data through on-site measurement or simulation using finite element simulation software. The sample data includes multiple groups of parameter data and construction results corresponding to each group. The parameter data includes time-varying parameters, material parameters, and environmental parameters. The second module is used to input the sample data into the initial model based on deep learning for processing in groups, and update the preset adjustable parameters in the initial model and the connection strength between each neuron based on the output of each neuron in each layer of the network in the initial model; The third module is used to calculate the loss function of each set of input data according to the output of the initial model based on the updated preset adjustable parameters and connection strength; A fourth module is configured to determine the initial model whose loss function reaches the preset training end condition as the crack assessment model until the loss function reaches the preset training end condition; The fifth module is used to obtain parameter data of the site to be evaluated, and input the parameter data of the site to be evaluated into the crack assessment model for processing to obtain the crack assessment results of the site to be evaluated, and the crack assessment results include high risk, medium risk and low risk.

9. An electronic device, characterized in that: The method comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the method according to any one of claims 1 to 7 is implemented when the processor executes the computer program.

10. A non-transitory computer-readable storage medium, characterized in that The non-transitory computer-readable storage medium stores computer instructions, which enable a computer to execute the method of any one of claims 1 to 7.

Images